For the batch crystallization process, the tracking of temperature trajectory, which involves severe? nonlinear and non-stationary features, is somewhat difficult to achieve. In this paper, three model-based control strategies, globally linearizing control, generic model control and multi-model based model predictive control, are applied to a batch crystallization process. Simulation results show that all three model-based controllers can provide superior trajectory tracking, disturbance rejection capability and robustness to the modelling errors in the process model as compared with conventional PI controller.