Chaining in NLP is a sequencing of a series of states.

An example would be chaining anchors. If you have set up a few anchors, you can fire them off one after the other, changing the state as each emotion is at its peak. You will then move through a sequence of states. A useful chaining anchor can make you (by firing it off a few times) go through the different states automatically, i.e. the first state will induce a process that automatically leads to the last state.