You really need to move this question to the discussion forum, so you can benefit from several people's answers. However here are my thoughts.
There are tons of great project ideas out there, but the one thing that makes a science fair project stand out is the scientific method. You should be trying to prove or disprove something with your project. Just building something, no matter how awesome it is, is not really science.
So start your project by using a statement like this: The effect of ___ on ____.
If you want to build something, that's fine, but then prove something with it. To use jackboniver's example of a pizza box oven, you could make one, set it up at different angles relative to the sun, and track the maximum temperature it reaches at each angle. Or you could build several different versions, changing one detail,from version to version (perhaps adding layers of insulation to the outside of the box?) and then set them all up and compare performance.