Abstract:To simulate the ecological land transition of Baotou City, Inner Mongolia, the MCR-ANN-CA model was built. This model was composed of minimum cumulative resistance model (MCR), artificial neural network (ANN) and cellular automata model (CA). The MCR model was used to simulate the resistance faced by ecological land during its transition. To measure the resistance, the data of normalized difference vegetation index (NDVI), digital elevation model (DEM), slope, distance from water, population density and industrial park distribution were normalized and superimposed to produce the resistance surface of MCR model. Then a cumulative consumption resistance surface was generated by using the cost distance tool of ArcGIS software. The generated surface was taken as the suitability map of CA model. The ANN model was used to extract CA neighborhood transfer rules and consider the neighborhood land use structure. The CA model combined the cumulative consumption resistance surface and the neighborhood transfer rules extracted by using the CNN model. And the MCR-CNN-CA model was finally formed. Based on the land use data of 2006 and 2011, the transition of construction land of Baotou City in 2016 was simulated by using the MCR-ANN-CA model. The simulation result of the model was compared with the CA-Markov model. The Kappa index of agreement (KIA) of the two models were 0.89 and 0.87, and the relative errors were 3.10% and 5.31%, respectively. The MCR-ANN-CA model showed high simulation precision.