For the two-step variety, you'll have to do a bit of light programming (examples here
http://forums.eviews.com/viewtopic.php?f=15&t=8744) to extract factors as per Stock and Watson (2002) (or otherwise), and then use your estimated factors in the second stage in a standard VAR model through the menus, or something like var favar.ls 1 2 x y z factor1 factor2 factor3 where x y z are your observed variables, and factor1/2/3 are your I(0) factors.
Any questions, please don't hesitate to ask.
Charlie