It could be argued that the science behind evolution is equally non-scientific as the science behind ID. If you think about it, replicating the primordial soup (and the pre-life world conditions), creating life from scratch, and waiting around for it to evolve in some way would take far too long to feasibly be possible. Does that make it not testable? Not falsifiable? If so, then it's not really science...
Fortunately, as discussed before, we can design experiments in a different way that test evolution without having to resort to recreating life from scratch. We can observe current organisms and past organisms and from these make falsifiable hypotheses. We can make models from our ideas and use them to make predictions about the future. If they don't correctly predict what happens, we can then look and see what's wrong to change the hypothesis.
This is what separates evolution science from ID science.