Understanding of the timing, orientation and transmission mode of crop disease spread is very important in prevention and treatment of crop diseases, especially for cross-border crop diseases. In this paper, we introduced the complex network approach to model trans-boundary spread of diseases. In the model, the process of crop disease spread is characterized as a network, where patches of crop planting are abstracted into nodes, and the diseases spread from a node to neighboring nodes. The edge between two nodes is connected, only if the Euclidean distance between two nodes does not exceed the radius of disease spread. We obtained wheat-growing distribution data of Beijing in 2007 by image classification, and further abstracted patches of wheat growing into network nodes, and finally constructed the disease spread network to analyze its topology characteristics. The experimental results show the heavy influence of spatial constraint on disease spread.