A great online course called "Writing in the Sciences" is now available free of charge, and it provides really great advices. When I was writing an article recently, I just had the time to watch the first week's lectures, and it helped me a lot to finish my manuscript.
To summary some quick tips to do a good writing:
1- Read papers and copy (everything but the content) from the ones you like. This will give you hints on how to structure a paper, style you can have and many other things. Mimicking is a fundamental learning mechanism of humans that helped us achieve the knowledge we have, don't overlook it.
2- Just get to write, even if you don't feel like it. A good paper is reworked several times before it's ready for publication, so just get going on your first pass, you'll enhance later.
3- Don't try to make overly long or complicated sentences. Just try to make short sentences, and try to use simple formulations (present tense instead of past tense, standard sentence form: subject-verb-object). Also avoid jargon and initials as much as possible, this will be one less burden for your reader (even when they're specialized in the domain you're writing in).
4- Be logical in your progression. Try to be logical and gradual from chapters to chapters (for this, generally there are common templates, like the OHERIC methodology: Observation/Problem introduction, Hypothesis/Your Model, Experiment/Simulations, Results/Interpretation, Conclusion/Opening) but also inside each chapter, so that the end of a chapter naturally leads to the next.
Usually, the introduction is the hardest part and usually too much overlooked so you should really focus on this part ; the conclusion is the easiest part (just summary what problem you tackled, what you did and what you've found and potentially future avenues that could extend the findings on this problem) ; and the abstract and title are done at the very end when you already wrote the whole paper.
5- Add lots of pertinent references. A good example is any well-developped article in Wikipedia (since this encyclopedia follows some common editorial guidelines that are just as well used in scientific publishing). For any claim you make that isn't yours, try to reference, and most importantly, your reference must be pertinent (avoid referencing an article you barely read or that isn't focusing at all on the claim you are making, ie: an article about brain's memory making an hypothesis about consciousness at the conclusion as an opening isn't fit to be referenced for any claim about consciousness since this is just speculation).
6- Be yourself. This can be difficult if you are not comfortable in the language you're writing your article in, but if you are comfortable enough, try to keep (or create) your own writing style.
7- I think this one is less important at first, but later if you want to be a pro: uniformize your editorial line. For example, if you use American English words in your abstract, don't use British English words suddenly in later parts. A very good listing of editorial tips can be found here:
8- Use a spellchecker, always.
9- Ask other people to read your paper, and be open to feedbacks. This is crucial and the final step to make a good paper, as it is very difficult to see the big picture and the small glitches yourself. If possible, try also to get your paper read by non-specialized persons, like your relatives, they will tell you if your work is pleasant to read even if they can't grasp every technical subtletlies.
Good luck for your paper. And remember: writing is always painful and feels unnatural for most people (including professional writers), so don't feel out of place, just try to do your job and try to be proud of the result.