In order to understand the urbanization process

logistic regression is introduced to mine the conversion rules of urban development. The rules are input into cellular automata (CA) model to simulate the urban development process of Foshan city. Experiments show that urbanization process is the core part in the development of Foshan city. The source of urban land comes from farmland

forests and other types of land (including orchards and ponds

etc.). In the simulation result

the overall accuracy and Kappa coefficient can reach 93.7% and 0.86

respectively. By using landscape indices

Foshan city will become more and more aggregative with the maximum patch index increase. It was found that there are 191 km

2

of farmland changed into urban land in the scenarios of urban prior development

accounting for 27% of all the farmland (706 km

2

). In the scenarios of farmland protection and forest protection

farmland and forest can be better protected

but urban expansion will occupy other types of land use to solve the problem of lacking enough land use for urban expansion. It is found that the spatial variables have remarkable corridor effect on urbanization by analyzing the relationship between proximity variables and urban land uses.