Sure, score voting is great for elections; but it's also good for any decision you need to make among more than two options. Right now, for example, some folks are worrying about nightmare scenarios where the major healthcare bill being discussed gets stuck in Senate limbo, because not enough senators would vote for the bill as it is (or rather, will be), but also, not enough senators would vote to change it toward the compromise that might allow the bill to proceed. (It's a similar problem to that of a legislative "poison pill" amendment.)
The problem is, the healthcare bill is big; it has lots of different parts, some of which are more popular than others, and all of which are supported by different groups of senators, and any one of them can be strengthened or weakened or changed entirely. So the overall choice the senate has to make isn't "yes or no," it's which set of pieces to use, and there are lot of different possible combinations. A lot more than two, at least.
It would (or at least, could) work like this: instead of a series of "yes or no" votes on changes to the bill followed by a "yes or no" vote on the final bill, a whole series of different bills could be considered at once (including a "none of the above" option), and each senator could give each of those bills a score in some range; the bill with the highest average score is the one that's passed (or if "none of the above" wins, no bill is passed.)
You might say "That's way too complicated!" but consider the alternative (i.e., consider what we do now): months of committee work to craft a bill that might have a chance on the floor, months more to reconcile the bills from different committees toward that same end, and weeks of debate on the floor, including dodging "poison-pill" amendments. So is it really more complicated? Just like in elections, is score voting really more complicated than dealing with our current ballot-access laws, party primaries, party conventions, and runoff elections? I think not.