It depends on what you mean by "correct".
Personally, I take it to be more a rule of thumb for making judgments about the utility of statements in ethics than a strict logical assertion relating to actions. In Critique, Kant writes that "the action to which the ‘ought’ applies must indeed be possible under natural conditions". In other words, we might say that a moral agent can be obliged to do something only if that action is within his capabilities of doing, both physically and in terms of whatever free will we might believe him to possess. So a more precise, albeit longer version, of ought implies can is perhaps if a valid moral judgement or command is made conerning an action, that action must be possible.
Is this correct for all such possible actions? Yes, trivially.
Is this correct for all (imagineable) actions? I'd say no. It is instructive to play around with the statement logically. Try to express the contrapositive (transform "a implies b" into "not b implies not a", which are logically equivalent). It will be can't imples shouldn't... does that make any sense? Not much in my mind. Treated literally, prohibiting something that is not possible seems meaningless. For example: you can tell a child that it ought not to fly around in the sky as a bird, but this will not have much influence on whether the child does takes your advice, since it is not within the child's abilities to fly at all. Because it has no choice.
On the other hand, there are valid oughts and invalid oughts and the consequences of their proclamation. That is, in the real world moral commands and judgements are actions in themselves, at least to the extent they influence other actions. The child may well try to fly when told not to, even though the shouldn't you dispensed is not rooted in a physically possible action. This can easily get complicated by what the child knows and believes to be true and how it reacts to our invalid ought. Is the essence of the prohibition (the ought) in our not wanting the child to fly, or in the child's understanding that we do not want it to try to fly? Who's ought is it? The same questions arise for valid oughts (you should not run near the pool), but the point here is that judgements made about impossible actions are real and affective. If we choose not do classify such judgements as real oughts, or widen the range of possible actions to all in principle possible actions including ones we can only imagine, then we return to the trivial case I started out with.
So, I think things eventually boil down to a consideration of the definition of possible actions, and therefore of the nature of human (free) will. From there it is only a step to how this is related (or not) to determinism, which has been written about a lot.
One interesting observation here is that if ethical statements are possible only for actions that are within the attainable realm (the strict sense), then increasing the possible also increases that domain on which we can make ethical statements. So, as the possibilities grow with technological progress, humans in that sense become "more ethical". I find that intriguing, and there is some discussion of this aspect in Dennett's book Freedom Evolves.