writing theses, papers, etc is an exponential process:
the first 50% takes 10% of the time, and the second 50% takes 90% of the time. The final 10% takes 50% of the time. ie. to make it perfect takes infinite time!
also, no matter how much you re-re-re-read your work, it will have a mistake. i'm sure the psychologists out there have a theory for this? i think it's called the 'law of sod'.
you can try re-reading your work after 5 days, re-reading it after a pint, and asking your super/colleagues to read it.