This chapter introduces the concepts of cellular automata (CA) which have been increasingly used for simulating urban dynamics. Simulation and prediction of urban evolution can provide the useful inputs to crime models. However, calibration of urban cellular automata is crucial for simulating realistic cities. Simulation of multiple land use changes using CA is difficult because numerous spatial variables and parameters have to be utilized. The incorporation of neural networks with CA can alleviate the calibration problems. This chapter illustrates how complex land use dynamics can be simulated by the integration of CA and neural networks.