Based on the advantages of single model in the field of land use simulation, This research construct a Markov-logistic-CA integrated model to reflect the land use change precisely considering both sides of macroscopic and microscopic factors. The conventional CA model is extended from two aspects including iteration times and conversion rules by Markov and Logistic respectively. Finally the integrated model software is developed by VC. NET language and tested in the shenzhen city. The experimental result shows that Markov-logistic-CA model can better simulate and predict the situation of land use change rationally, and the result reached the expectant aim from the model accuracy test. It is conclusion that the Markov-logistic-CA model is a good tool to make decision for local government and have powerful computation function in complicated geographical system.