Partially observable Markov decision processes (POMDPs) have gained signicant interest in research on spoken dialogue systems, due to among many benets its ability to naturally model the dialogue strategy selection problem under the unreliability in automated speech recognition. However, the POMDP approaches are essentially model-based, and as a result, the dialogue strategy computed from POMDP is subject to the correctness of the model. In this paper, we extend some of the previous user models for POMDPs, and evaluate the effects of user models on the dialogue strategy computed from POMDP.