Review the material at hand. Decide what your instincts tell you about what you've observed and what you believe might happen.
Look for clues using common sense or past scientific material. Find something you can relate to the current situation that will allow you to form a reasonable relationship and prediction in your own matter.
Make a list of possible hypotheses that could be tested. The ultimate goal of a hypothesis should be to create something that is entirely provable or disprovable based on supporting evidence and tests. If you can't put your hypothesis to work its useless in the realm of science.
Select a hypothesis you believe can easily be tested and serves a purpose to further scientific inquiry. Don't choose something too obvious or too difficult. Go with your empirical instinct based on what you've seen before.
Write out the hypothesis in a clear and straight forward manner. It should be obvious to anyone what you're proposing.