Maybe if you're being very broad in definitions then some class of proposals describable as "dark matter" might be unfalsifiable, but to be taken seriously as a scientific proposal I think it should be specific, concrete, and indeed testable, and there are a few of these within the "dark matter" class.
Again, we're in the perhaps unsatisfying position of having observations which don't cohere with our current theoretical understanding. What's the solution? It's not easy...