Abstract: In order to improve the stability of the train-wayside wireless communication of urban rail transit, a trainwayside wireless communication scheme based on multi-network integration was proposed in this paper which used a variety of wireless networks to ensure data transmission. Selecting the optimal network for data transmission was a key technical point in this scheme. The paper proposed an optimization network selection algorithm based on support vector machine(SVM) to solve network selection problems by classification method. Then experimental simulations were used to prove the availability and reliability of SVM on optimal network selection. In this paper, the simulation environment of multi-network integrated train-wayside communication was established. the simulation results show that the train can intelligently detect and switch to a better quality wireless network to complete train-wayside data communication when the signal quality of the current wireless network was degraded. The train-wayside multi-network integration wireless communication system scheme based on SVM can improve the wireless communication quality of train-wayside communication. The scheme has the effectiveness and feasibility.