21 STEM Project Ideas for High School Students

21 STEM project ideas and research topics for high school students, from biology to math, with the method, materials, level, result and approvals for each.

Written by
Last updated
October 31, 2026
16
minute read
Project Ideas
Tasmir Aziz
Tasmir Aziz
M.S., Oxford University
Tasmir Aziz is a Contributing Writer at Nova Scholar. He holds an M.S. from Oxford University and has mentored 100+ students on college admissions, helping them develop research ideas, craft compelling personal narratives, and refine application essays.
About the author

We've found that good STEM projects for high school students answer one question with data you collect or download yourself. Below are 21 ideas in biology, chemistry, physics, engineering, computer science, environmental science and math. Each lists the method, materials, level, result and any approval you need.

We picked ideas that take 1 to 6 weeks, need little or no lab equipment and end in numbers you can check. For writing, art or service projects, see our passion project ideas for high school students.

What STEM projects can you do in biology?

These biology projects use pineapple enzymes, public protein data and your own heart rate, so you need a kitchen, a laptop or a fitness tracker. Only the heart-rate idea needs approval, because it measures a person.

For more, from planarian regrowth to coral heat data, see our biology project ideas for high school students.

1. Pineapple enzymes and gelatin

  • Question: At what temperature does heat stop pineapple's bromelain enzymes, which break down proteins, from ruining gelatin?
  • What you do: heat fresh juice to six temperatures for five minutes each, cool it and stir it into lukewarm gelatin, three cups per temperature. Add unheated-juice and plain-water cups, and see which set.
  • You need: a fresh pineapple, unflavored gelatin, a pot, a cooking thermometer and cups
  • Level: beginner; 1 to 2 weeks
  • You'll have: the temperature where the enzymes stop working
  • Approval: none; heat the juice with an adult nearby

2. One protein across species

  • Question: Is cytochrome c less like the human version in species that split from us longer ago?
  • What you do: compare the protein in 10 to 15 species with the human version, then plot similarity against split time.
  • You need: NCBI's Protein BLAST and TimeTree, a public database of when species diverged
  • Level: intermediate; 3 to 4 weeks
  • You'll have: a scatter plot of percent identity against split time, and the species that don't fit
  • Approval: none

3. Wrist tracker accuracy

  • Question: How closely does a wrist tracker's heart rate match a chest strap at rest, walking and jogging?
  • What you do: wear both devices and record paired readings every minute on at least five days, treating each day as one trial.
  • You need: a wrist fitness tracker, a chest-strap heart rate monitor and a track or treadmill
  • Level: intermediate; 4 to 6 weeks, plus review time
  • You'll have: the average gap between the devices at each activity level, for your own wrist
  • Approval: for an ISEF-affiliated fair, IRB approval before any data, even as the only participant, because ISEF's rules count physical activity as human participant research. See our guide to getting IRB approval.

What are good chemistry project ideas for high school students?

These chemistry projects use grocery and pharmacy supplies and take 1 to 3 weeks each. None needs review-committee approval. But ISEF requires a risk assessment with a supervising adult before you use hazardous chemicals like the iodine and rubbing alcohol below, even at home.

For more experiments, such as testing sunscreen, see our science fair project ideas.

4. Vitamin C in stored juice

  • Question: How fast does vitamin C break down in orange juice that's chilled, left out or boiled?
  • What you do: count the drops of iodine it takes to turn juice and starch blue-black, using a dissolved vitamin C tablet as your standard.
  • You need: orange juice, vitamin C tablets, iodine tincture, cornstarch, droppers and safety goggles
  • Level: intermediate; 2 to 3 weeks
  • You'll have: vitamin C per 100 mL over seven days; check the fresh reading against USDA's FoodData Central
  • Approval: none; do the risk assessment with an adult using the iodine's Safety Data Sheet. Never taste samples, and discard any that grow mold.

5. Ink chromatography

  • Question: Can paper chromatography tell which black pen wrote a note?
  • What you do: dot each pen's ink on paper strips, stand them in water or rubbing alcohol and measure how far each dye travels. Then identify the pen behind a mystery note.
  • You need: five or more black pens, coffee filters, water, 70% isopropyl alcohol, jars and a ruler
  • Level: beginner; 1 to 2 weeks
  • You'll have: an Rf value (dye distance ÷ liquid distance, both from the starting line) for each dye, plus your blind-test result
  • Approval: none; rubbing alcohol is flammable, so do the risk assessment and keep it away from flames

6. Crystal size and cooling speed

  • Question: Does slower cooling grow larger alum crystals?
  • What you do: pour one hot alum solution into nine jars and cool three each in a fridge, on a counter and in a cooler.
  • You need: alum from the grocery spice aisle, distilled water, jars, a thermometer, a small cooler and calipers
  • Level: beginner; 2 to 3 weeks
  • You'll have: a cooling curve for each setting and the average size of the five largest crystals in each jar
  • Approval: none; read alum's Safety Data Sheet and use heat-proof gloves

What are good physics project ideas for high school students?

Our advice: pick a physics project that tests a prediction from a formula, so you can compare what you measure with what should happen. The marble and magnet ideas run on a phone and free apps.

7. Marble energy loss

  • Question: Does a marble lose a bigger share of its energy when it starts higher?
  • What you do: release it down a U-shaped foam track from four heights and record how high it climbs on the far side.
  • You need: foam pipe insulation cut lengthwise, marbles, a meter stick, a phone with slow-motion video and the free Tracker video tool
  • Level: beginner; 2 weeks
  • You'll have: the share of energy kept at each height. The worked example below shows the plan.
  • Approval: none; cut the foam on a board, blade moving away from your hand

8. Laser bending in sugar water

  • Question: How does sugar concentration change how much a laser beam bends as it enters water?
  • What you do: shine a laser into sugar water at 60° from a line perpendicular to the surface. Measure the bent beam's angle from the same line and apply Snell's law.
  • You need: a clear box-shaped container, a laser pointer, a printed protractor, sugar and a kitchen scale
  • Level: intermediate; 2 to 3 weeks
  • You'll have: a graph of refractive index against sugar concentration
  • Approval: none; follow the FDA's laser advice: 5 mW or less, never aimed at a person or animal. Watch for reflections off the glass and water.

9. Magnet strength and distance

  • Question: Does a magnet's field drop with the cube of distance, as a university physics textbook predicts far from the magnet?
  • What you do: find your phone's magnetic sensor and move a small magnet straight away from it in 1 cm steps, pointed the same way. Start where the reading isn't maxed out and subtract the no-magnet reading on each axis.
  • You need: a phone, the free phyphox app, a small disc magnet and a ruler
  • Level: intermediate; 1 to 2 weeks
  • You'll have: a log-log graph of field against distance, with its slope compared with -3
  • Approval: none

What are good engineering project and prototype ideas?

Each engineering idea here ends with a working prototype plus test data showing which design works best. Our advice: build at least two versions, since one prototype can't show which choice made it work.

For more builds, see our guides to engineering projects for high school students, mechanical engineering project ideas and robotics and drone projects.

10. 3D-print strength

  • Question: Which infill pattern and density make the strongest 3D-printed hook per gram of plastic?
  • What you do: print one hook design at six settings, three copies each, then weigh each hook and load it until it breaks.
  • You need: a school or library 3D printer, PLA filament, a bucket of sand and a luggage scale
  • Level: intermediate; 3 to 4 weeks
  • You'll have: breaking load per gram for each setting, with photos of each break
  • Approval: none; wear safety glasses (PLA snaps into sharp pieces) and keep your feet clear of the bucket

11. Solar car gearing

  • Question: Which gear ratio gets a small solar car across 5 meters fastest?
  • What you do: swap three or four gear ratios on one car and time five runs of each in midday sun.
  • You need: a solar car kit with spare gears, a stopwatch and a light meter app
  • Level: beginner; 2 to 4 weeks
  • You'll have: the average run time for each ratio, with light readings to flag runs slowed by clouds
  • Approval: none

12. Cool roofs on model houses

  • Question: How much cooler does a white or planted roof keep a model house than a black one?
  • What you do: build three identical foam-board houses with different roofs and log their inside temperatures every 15 minutes in sun. Swap their spots daily.
  • You need: foam board, black and white paint, a sedum or grass mat and three probe thermometers
  • Background: according to the EPA, cool roofs can lower peak indoor temperatures by 2.2 to 5.9°F in homes without air conditioning
  • Level: beginner; 3 to 4 weeks
  • You'll have: each house's daily peak temperature over five or more sunny days, compared with EPA's range
  • Approval: none; use a craft knife on a cutting mat

What are good computer science STEM projects?

Each computer science idea here measures something you can graph, using free tools and public data. The spam filter and traffic model need some Python or NetLogo coding; the QR test needs none.

For AI builds, see our AI project ideas for high school students.

13. QR code damage

  • Question: How much of a QR code can you cover before it stops scanning, at each error-correction level?
  • What you do: print one code at levels L, M, Q and H, cover more of it step by step and scan with two phones. Keep the three corner squares clear, since scanners use them to find the code.
  • You need: a free QR generator that lets you pick the level, a printer and sticky notes
  • Level: beginner; 1 to 2 weeks
  • You'll have: the share covered when each level fails, compared with the restore rates from DENSO WAVE, the QR code's developer (Level M: about 15%). Those rates count damaged data, not covered area.
  • Approval: none

14. A spam filter from scratch

  • Question: Can a spam filter you write yourself match a ready-made one?
  • What you do: code a naive Bayes filter in Python, then test it and scikit-learn's version on the same messages.
  • You need: Python and UC Irvine's free SMS Spam Collection of 5,574 labeled texts
  • Level: intermediate; 3 to 5 weeks
  • You'll have: precision and recall for both filters, plus the messages each one got wrong
  • Approval: none, since ISEF exempts public datasets

15. Phantom traffic jams

  • Question: What share of gently braking cars stops phantom traffic jams in a simulation?
  • What you do: edit the Traffic Basic model so some cars brake gently, then run 10 trials at each share from 0% to 50%.
  • You need: Northwestern University's model, which shows jams forming without accidents, in NetLogo Web
  • Level: advanced; 4 to 6 weeks
  • You'll have: average speed and the number of stopped cars at each share, on one graph
  • Approval: none

What are good environmental science project ideas?

One of these projects uses mealworms, live invertebrates that ISEF's animal rules don't cover. Your school or fair may still set its own rules, so ask your teacher before you order them, and never release them outdoors. The other two use a nearby stream and federal electricity data.

16. Road salt in a local stream

  • Question: How much does a salted road raise chloride in a stream it crosses, and is the rise biggest after a thaw?
  • What you do: measure conductivity just above and below the road on at least eight winter days, including some after a thaw. Use the same two spots each time, treat each day as one trial and average three readings per spot. Then measure five salt mixes of known strength, made with distilled water, to turn readings into salt levels.
  • You need: a pocket conductivity meter with temperature compensation, table salt, distilled water, a kitchen scale, measuring cups and a cup on a pole. Our estimate: $20 to $40 if you own the scale.
  • Background: the EPA's 1988 freshwater criteria for chloride are 860 mg/L for brief exposure and 230 mg/L for continuous exposure. Sodium chloride is about 61% chloride by weight, so convert first.
  • Level: intermediate; 4 to 6 weeks
  • You'll have: a graph of chloride above and below the road on each day, beside EPA's numbers. Your figures may run high, since conductivity tracks all dissolved minerals together.
  • Approval: none under ISEF rules, since it involves no people, animals, microbe cultures or hazardous chemicals. Go with an adult, sample from the bank, stay off ice and wash your hands afterward.

17. Mealworms and Styrofoam

  • Question: How much Styrofoam do mealworms eat, and do they grow as well as bran-fed ones?
  • What you do: keep two bins of 50 on each diet: foam only, foam plus bran, and bran only. Weigh the foam every few days for four weeks, including foam in a worm-free bin that shows humidity changes.
  • You need: pet-store mealworms, clean plain foam, wheat bran, ventilated bins and a scale that reads to 0.001 g
  • Background: in a Stanford study, 100 mealworms ate 34 to 39 mg of Styrofoam a day, too little for a kitchen scale
  • Level: beginner; 4 to 5 weeks
  • You'll have: foam eaten per 100 mealworms per day, plus survival and weight gain on each diet
  • Approval: none under ISEF rules

18. Your state's electricity mix

  • Question: How have coal, gas, wind and solar shares of your state's electricity changed since 1990?
  • What you do: calculate each source's share by year, then look for plants added or retired around the biggest change.
  • You need: the EIA's state generation and capacity spreadsheets, which cover 1990 to 2025 as of October 2026
  • Level: beginner; 2 to 3 weeks
  • You'll have: a stacked area chart and a two-page explanation of the largest shift
  • Approval: none

What are good math project ideas for high school students?

A math project becomes research when you test a rule against real data or a big computation. Only the Benford idea works without coding; the other two need basic Python.

19. Breaking ciphers

  • Question: How long must a message be before letter counts crack a Caesar or Vigenère cipher almost every time?
  • What you do: write code that encrypts passages of 20 to 1,000 letters, then cracks them by matching English letter frequencies.
  • You need: Python and free eBooks from Project Gutenberg
  • Level: intermediate for Caesar, advanced for Vigenère; 3 to 6 weeks
  • You'll have: success rate against message length for each cipher, and the length that works 95% of the time
  • Approval: none

20. Benford's law in county populations

  • Question: Do the first digits of US county populations follow Benford's law?
  • What you do: count how often each first digit appears and compare with Benford's prediction, log10(1 + 1/d), using a chi-square test.
  • You need: the Census Bureau's county population estimates (Vintage 2025, the latest as of October 2026) and a spreadsheet
  • Level: beginner; 1 to 2 weeks
  • You'll have: a bar chart of observed and predicted shares, plus a test on data you expect to fail
  • Approval: none

21. Gaps between primes

  • Question: Does the average gap between primes near n grow like ln n, as the prime number theorem implies?
  • What you do: write a sieve that lists every prime up to at least 10 million, then average the gaps in blocks.
  • You need: Python or another language and a laptop
  • Level: intermediate; 2 to 4 weeks
  • You'll have: a graph of average gap against ln n and a table of the largest gaps you found
  • Approval: none

Which STEM project fits your grade and schedule?

Match the project to the weeks you have and the skills you already use. A first project should be short and cheap, while later ones can add code or an approval step. Our advice: finish one idea before you start another.

  • Middle school or 9th grade: ink chromatography (5), the solar car (11) or QR codes (13). See our passion projects for middle school students and engineering projects for middle school students.
  • 10th grade: test a formula with marble energy loss (7), magnet distance (9) or Benford's law (20). If you already code, the spam filter (14) is a stronger pick.
  • 11th grade, aiming for a spring fair: start in early fall. The wrist tracker (3) needs IRB review before any data, so look up your fair's approval deadline first.
  • 12th grade, during applications: pick something you can finish in a month, such as the electricity mix (18) or the protein comparison (2).
  • No lab or budget: ideas 2, 14, 15, 18, 19, 20 and 21 need only a laptop.

How do you turn a STEM research topic into a project you can finish?

Narrow the topic to one thing you change and one thing you measure, then plan a comparison and at least three runs per condition. ISEF's rules say your research plan should be completed before experimentation. Here's how "the physics of roller coasters" became the question in idea 7.

Worked example: from roller coasters to idea 7

  • What you change: release height: 20, 40, 60 and 80 cm, measured straight up from the bottom of the U
  • What you measure: how high the marble climbs on the far side, measured the same way (build that side taller than 80 cm)
  • What stays the same: the marble, the track and how you let go
  • Repeats: five releases per height, 20 in all. Re-measuring one video doesn't count as another trial.
  • The math: at its highest point the marble stops for an instant, so all its remaining energy is potential energy. Mass and gravity cancel, leaving energy kept = climb height ÷ release height.
  • For example: a marble released from 60 cm that climbs to 51 cm kept 85% of its energy. Rolling friction in the foam, air resistance and sound took the rest.
  • Reading the result: average the five runs at each height and plot energy kept against release height. Our advice: call it a trend only if it's bigger than the spread between runs at one height.

Comparing heights avoids a common error: a rolling marble stores some energy in its spin, so its speed alone undercounts its energy.

In our Nova Research Program, you work one-on-one with a mentor to narrow a topic into a research question and write a paper.

Which STEM projects need approval before you start?

Under ISEF rules, projects with human participants, vertebrate animals or potentially hazardous biological agents need review-committee approval before you start. In this list, only the wrist tracker (3) does, while ideas 4 and 5 need a risk assessment. ISEF says affiliated fairs may add restrictions, and your school can be stricter, so ask your teacher before you buy supplies.

  • People: an IRB approves the project before any interaction, even when you're your own subject. Testing your own invention or a store-bought product alone is exempt if it's safe. Adding a human variable, such as exercise or sleep, ends the exemption.
  • Vertebrate animals: your fair's Scientific Review Committee (SRC) approves the project first. The rules cover all nonhuman vertebrates, including fish, plus cephalopods such as octopuses. Only watching animals, with no contact and no change to their surroundings, is exempt.
  • Microbes, recombinant DNA, tissue and body fluids: SRC approval and a lab rated BSL-1 or higher. But ISEF's body fluid rules let you study your own uncultured fluids at home; ask your fair's SRC first. Body fluid samples that can be identified with a specific person need IRB approval and informed consent. ISEF prohibits culturing microbes at home, even low-risk BSL-1 ones, and unknown microbes from soil or skin count as BSL-2 once you open the dish. A few studies, such as baker's yeast fermentation, are exempt but still need a risk assessment.

Rules checked October 30, 2026 on Society for Science's ISEF rules pages, last updated June 2026.

Guides in this series

Frequently asked questions

What is the difference between a STEM activity and a STEM project?

An activity usually fits in one class period, like building a paper bridge. A project runs for weeks around one question and ends with data, a prototype or a paper. Ink chromatography (5) works as a one-day activity if you skip the blind test.

How do you make a STEM project unique?

Change one thing about a familiar idea: use local samples, add a variable or compare your result with an official figure. Ideas 12 and 16 do this by setting your data beside EPA figures. We'd pick a familiar question answered with your own data over a novel idea you can't finish.

How long does a STEM project take?

Each idea lists its own time; add time to plan, write and, for idea 3, get IRB approval. An ISEF-affiliated fair allows no more than 12 months of continuous research. For the 2027 fair, work done before January 2026 doesn't count.

Where can you enter a finished STEM project?

Start with your school or regional science fair. Our guide to science fairs and competitions for high school students compares the main options, with 2026-27 deadlines and how to reach Regeneron ISEF.

Ready to turn potential into real achievements?

Discover the right pathway, learn about timelines, and see how our mentorship-driven approach can help your student succeed.
Download the Programs Brochure