Know Your Audience.
Assume that the grant reviewers are peers who are generally familiar with your field, but not with the specifics of your research topic and problem.
Write a Complete Proposal.
You should summarize the issue your study addresses, background about the problem, what you want to do (specific aims), the idea that you want to test (your hypothesis or goals), preliminary data to support the hypothesis (if applicable), a practical experimental design and methods to test the hypothesis or meet the goals, possible immediate outcomes that you predict from your approach, your plan to manage potential pitfalls, and the long term outcomes (significance) that you predict will impact the field.
Sell your proposal by stating its significance and the feasibility of your approach, including realistic time/budget constraints.
Think like a reviewer. A reviewer must read approximately 20 applications and form an opinion about each one. Organize the proposal logically and clearly. Your thought process of the application should be easy to follow. Make your points as directly as possible.
Write in plain language so a non-expert may understand your proposed work.
Include enough background information to enable a reviewer to understand your proposed work, but keep it succinct so you don't exceed the page limits.
Note that page and word limits are firm maximums; fewer pages are encouraged if the necessary information can be provided in less space. Please observe the formatting requirements on page 2 of the application and the page limits for each section. Applications that exceed these limits will be disqualified.