My only concern about using nuclear charges on potential impactors is based on the composition of the object. I could see many situations where simply firing a charge on the object's surface could simply fragment it, leaving the large chunks still on a collision course and no viable means to redirect the many new objects. This might be particularly true of cometary objects. Would chunks (nice technical term) over a wider area create less overall global damage than one large object? Perhaps. I don't know. Maybe with sufficient warning test probes could first be sent to examine the structure of the object to see if a direct nuke would work.

Another approach would be using some form of propulsive device to "land" on the object, and slowly change its course with a thrusting device (a mass driver using pieces of the object as thrust mass would be the most effective, but far more complicated to develop and deliver than a simple nuclear charge).

A compromise might be to develop an "Orion"-type thrust system using many small nuclear charges attached to a shock-absorbing unit connected to the asteroid or other object, preventing the charges from fracturing the object if it's fragile.