I don't identify as an incompatibilist, but to try to steelman their position:
Some incompatiblist determinists might view metaphysical libertarianism as a logically coherent possibility, albeit a false one. They might then view "free will" as useful in naming a way the universe coherently could have been, but turned out to actually not be – much like they might view Ptolemaic epicycles, phlogiston, the luminiferous aether, and steady state theory. And, much as a person might object to the position "the luminiferous aether is actually true, if we redefine all the terms involved to refer to concepts from special and general relativity" as a form of unhelpful obscurantism, they might view compatibilist determinism as doing the same kind of thing to metaphysically libertarian free will.
There's also the other kind of incompatibilism: metaphysical libertarianism is itself a form of incompatibilism. Both incompatibilist determinists and metaphysical libertarians agree that determinism and free will are incompatible. They just disagree on whether that means we should junk free will or junk determinism.
Although I think here "determinism" is a bit of a misnomer. Most physical determinists don't actually have a problem with the idea that there might be irreducible quantum indeterminism (whether or not they believe there actually is). An incompatibilist determinist would say "a clock doesn't have free will, and a roulette wheel doesn't either, so neither can some hybrid between the two". A compatibilist determinist will say that whether it is a clock or a roulette wheel or a hybrid of them, if it gets sufficiently complicated in the right ways, then it will have free will. For both, the real objection is to metaphysically libertarian ideas that human choice involves some irreducible "third thing" which is neither deterministic causation nor impersonal chance, nor any mere combination of the two. There are also claims (e.g. by Roger Penrose) that quantum indeterminism operates in some special way within the human brain, different from how it operates normally, and that human free will is somehow rooted in that. Many "determinists" (whether compatibilist or incompatibilist) would view that as uncomfortably close to metaphysical libertarianism, even if it does attempt to partly replace the transcendental metaphysics with something closer to the realm of the empirically testable (even if not quite there). But, while "quantum indeterminism operates in a very special way in the human brain" makes them uncomfortable, I don't think most of them view in the same way "quantum indeterminism is real and irreducible, but it doesn't operate in the human brain in any way differently from how it operates anywhere else"–again, whether or not they actually believe it. Some will prefer purely deterministic interpretations of quantum theory instead, such as many worlds or de Broglie–Bohm theory–but that debate has no inherent connection to the topic of human free will.
Some incompatiblist determinists might view metaphysical libertarianism as a logically coherent possibility, albeit a false one. They might then view "free will" as useful in naming a way the universe coherently could have been, but turned out to actually not be – much like they might view Ptolemaic epicycles, phlogiston, the luminiferous aether, and steady state theory. And, much as a person might object to the position "the luminiferous aether is actually true, if we redefine all the terms involved to refer to concepts from special and general relativity" as a form of unhelpful obscurantism, they might view compatibilist determinism as doing the same kind of thing to metaphysically libertarian free will.
There's also the other kind of incompatibilism: metaphysical libertarianism is itself a form of incompatibilism. Both incompatibilist determinists and metaphysical libertarians agree that determinism and free will are incompatible. They just disagree on whether that means we should junk free will or junk determinism.
Although I think here "determinism" is a bit of a misnomer. Most physical determinists don't actually have a problem with the idea that there might be irreducible quantum indeterminism (whether or not they believe there actually is). An incompatibilist determinist would say "a clock doesn't have free will, and a roulette wheel doesn't either, so neither can some hybrid between the two". A compatibilist determinist will say that whether it is a clock or a roulette wheel or a hybrid of them, if it gets sufficiently complicated in the right ways, then it will have free will. For both, the real objection is to metaphysically libertarian ideas that human choice involves some irreducible "third thing" which is neither deterministic causation nor impersonal chance, nor any mere combination of the two. There are also claims (e.g. by Roger Penrose) that quantum indeterminism operates in some special way within the human brain, different from how it operates normally, and that human free will is somehow rooted in that. Many "determinists" (whether compatibilist or incompatibilist) would view that as uncomfortably close to metaphysical libertarianism, even if it does attempt to partly replace the transcendental metaphysics with something closer to the realm of the empirically testable (even if not quite there). But, while "quantum indeterminism operates in a very special way in the human brain" makes them uncomfortable, I don't think most of them view in the same way "quantum indeterminism is real and irreducible, but it doesn't operate in the human brain in any way differently from how it operates anywhere else"–again, whether or not they actually believe it. Some will prefer purely deterministic interpretations of quantum theory instead, such as many worlds or de Broglie–Bohm theory–but that debate has no inherent connection to the topic of human free will.