First you need to come up with a stack of references. It always looks good to reviewers if you've got a reference list as long as your arm. You need a small army of other scientists on whose broad shoulders your authority may rest. It looks bad if each and every idea, no matter how trivial, is not at least hinted at in some other work. Heaven forbid you should come up with something original.
Always pepper your writing with plenty of faux Latin phrases. Terms like "in situ" "et al. "i.e." "vis a vis" "a priori" "post hoc" etc. It makes you sound educated.
Whenever you come up with a new method, algorithm or computer program, Christen it with a cute, catchy or flashy acronym. ARTS (Atmospheric Radiative Transfer Simulator) and ARTIST (I forget what this one stands for but it's for sea ice) are two nice ones from my own field. It's also good to abbreviate every technical phrase you use more than twice. It makes you look like you've got a firm, insider grasp of topics that no one else understands.
Try to squeeze in as many equations as possible. Concepts that are easy to explain in words can always be made more obtuse by converting them into formulas, er, sorry, formulae. Besides, like a long reference list, it looks impressive. If your reviewers don't have to take at least 3 Tylenol IIIs (six regular) after puzzling unsuccessfully over your article, it just isn't ready for prime time.