As I remember it, Hume argues that the evidence for any miracle - defined as a violation of the laws of nature - is always going to be less persuasive than the evidence for the alternatives (illusion, delusion, deception, wishful thinking, etc.). All the latter are commonplace, whereas miracles are by definition rare, so why wouldn't we believe the more obvious, common and therefore likely explanation of events?
That's a good argument for disbelieving each and every individual miracle, and since the entirety of the set of miracles is made up of all the individual instances we might infer that it's an argument against believing miracles altogether. But wait - said my feverish brain. Supposing that when Napoleon lost Waterloo, instead of sending him to St Helena they had clapped him in an asylum for the insane, along with 99 lookalike prisoners, all of whom were suffering from the delusion that they were Napoleon? Suppose that all these prisoners had obsessively researched Napoleon's past and character, and knew enough about him to be as convincing in the part as the general himself? Any visitor to the asylum would have excellent reason to disbelieve the protestations of each and every prisoner that they were the former emperor of France - after all, it would be 99/1 against - but nevertheless, one of them really would be Napoleon. Here - and surely also with miracles - an argument against accepting any and every individual instance of a class is not an argument that there are no instances of a class, and indeed it may be compatible with believing that there are such instances.
Possibly Hume actually anticipated all this: I can't remember, and I'm not going to check right now. But he's frequently cited as arguing against the existence of miracles in general, and by that I'm unconvinced.
Having said all that, I don't believe in miracles, at least as Hume defines them, because I'm not convinced by the whole "laws of nature" thing, which seems an unwarranted literalization of a metaphor borrowed from human jurisprudence.