Trivially yes, and [see edit!] even along any axis/direction.
There are three axes along which a muscle could be "extended". Let's say the muscle is aligned with the z-axis (so, by contracting, it would pull something along the z-axis). Then we could extend the muscle along the z-axis (making it longer), the x-axis (making it wider), or the y-axis (making it...taller, I suppose.)
I'm not certain what would happen if you extend the muscle along the z-axis. I expect a better answer than mine will come along and delve into that.
But if you were to extend the muscle along the x or y axis, you would certainly increase its power, linearly with the amount it has been extended. Consider a detached bicep muscle, floating in space. If it contracts, it would exert some power P. Now imagine an identical bicep muscle beside it. If both were to contract, they would exert a power of 2P. Now squish them together, so that they're effectively one muscle that's twice as wide (ie. extended along whichever single axis corresponds to "width"). You now have a single muscle, which has been scaled by a factor of 2 along one axis, that would contract with power 2P.
Edit: It's time to deal with that question of attachment sites!
(Inspired partly by reading Demigan's excellent comment on my answer)
So you can totally increase power along the z-axis, if you're willing to play with attachment sites. Let's take our two floating biceps and arrange them end-to-end along the z-axis. If we connected them to each other, in series, then we increase the range-of-motion but not the power. If, however, each bicep runs a tendon past the other, bypassing it and connecting directly to the target attachment site, then we have two muscles arranged as if in series but actually connected in parallel. This arrangement would increase the power but not the range-of-motion--the double-muscle could only move its load up to the same distance as a single muscle.
(As described, though, this is a hack. For starters, we'd probably want the muscles to be running tendons through each other, so that they're not awkwardly pushing each other to the side when contracting. And I'm sure there would be even better ergonomic arrangements than just having the two muscles stacked end-to-end.)