The 'actual world' -- better, the actual state, or history of the world -- should not be confused with the enormous scattered object that surrounds us. The latter might also have been called 'the (actual) world', but it is not the relevant object here. Thus the possible but not actual worlds are not phantom duplicates of the 'world' in this other sense. Perhaps such confusions would have been less likely but for the terminological accident that 'possible worlds' rather than 'possible states', or 'histories', of the world, or 'counterfactual situations' had been used. (pp.19-20)
Note that what he calls the 'actual world' is what I would call the possible world/scenario corresponding to our concrete world. But despite the terminological difference, the key point is the same: this correspondence is not identity.