Don't be so hard on yourself. The success rate for most significant research grants/fellowships is very small, you should expect to be rejected for most grants you apply for, it's just the nature of the game unfortunately.
You don't mention what country you are working in and what granting bodies you are applying to. Each has it's own criteria and approach. There are many books and other resources which attempt to teach grant writing. Most universities offer some kind of grant writing workshops/support. The ones I have read/workshops I have attended I didn't find particularly useful.
The following are some suggestions which I think are important (in no particular order):
- Get feedback on your application. Both from people who have been successful writing grants as well as from people outside your field. Keep in mind the committee is unlikely to have specialists working in your exact area of research - your case must be compelling to a broad range of researchers. You need to find people who will be brutally honest with you not people who will be encouraging.
- Include some preliminary results to strengthen your proposal.
- Try to get feedback from the grant bodies if they provide it.
- You need to convincingly explain what makes you/your project special, why are you the best/only person that can do it? Why is your university the best/only place you can do this work?
- Get to know the people who are on the committees for deciding grants (ie. senior academics). From my experience you need someone on the committee to champion your application - the more friends in high places you have the better. Avoid making enemies - even a single reviewer can kill your chances.
- Get some experience reviewing grants/proposals (even small internal ones) this will really help you to write grants considering the committee's perspective.
- Read the call very carefully and address every aspect clearly.
From my experience reviewing grants many are far too light on specific details. Eg. exactly what methods will you use (how many test specimens, what loading, what are the interesting parameter ranges, etc)? Why will your approach work and what is the risk mitigation strategy? Etc
Ultimately the more you write the better you will get at it.
