I've recommended that a scientific presentation should be based on a concise written narrative (1/2 page or so, 1-3 minutes orally) that qualitatively describes the major points to be made in the presentation. I've suggested that scientists try to write out this narrative before making slides.
The problem is that it's not easy to write a narrative and many people don't feel confident about their writing. So if you have trouble creating a narrative, here's a few tips:
1) How many slides do you remember 24 hours after hearing a talk? Not many, if any. So when a member of your audience wakes up the next morning, what major points (2-4) do you hope they remember? Write these out and then try writing a narrative that leads to these insights.
2) Alternatively, if you have an abstract from a journal article, start with that. Try to make it more conversational, as if you were explaining it to talk about it with somebody. Will you be able to cover all the points in contained in the abstract when you present, or should your narrative short circuit a few?
3) Alternatively and less good is to start with specific aims from a proposal. The problem is that specific aims usually look forward and don't always provide the necessary background that lets the audience understand the importance of the aims. Again try to make it more conversational.
4) Try making an outline for your talk. Hopefully the talk can fit into three or four sections. Identify the major points you want to make in each section and then try to string together a written narrative. Or simply write a narrative for each section.
5) Always give an oral version (hopefully 1-3 minutes) of your narrative to colleagues (hopefully a colleague in your subspecialty as well as one who is a bit removed from your area). It is just as important to rehearse this narrative as it is to dry run your slides. Without a coherent narrative, you don't have a presentation, you only have a data dump.