The transportation system has entered the era of 5G intelligent Internet of things (IoT), which can realize the comprehensive monitoring, perception, and intelligent decision-making of people, vehicles, roads, and the environment. The purpose is to solve the problems in the communication security of intelligent transportation system (ITS) and improve the vulnerability of traditional distributed architecture. The security issues of the Internet of vehicles (IoV) in 5G environment are analyzed from the perspective of big data. An access control mechanism based on risk prediction is proposed aiming at the problems existing in the node access control process. A Wasserstein Distance-based Combined Generative Adversarial Network (WCGAN) is proposed. It modifies the loss function to solve the gradient disappearance problem, and a combination of multiple generators is designed to solve the pattern collapse. The simulation experiment is carried out on the dataset of the intrusion detection evaluation project. The WCGAN model has the smallest prediction error than the other models regarding the node packet transmission rate. Its loss value is close to 0 after 10 iterations, while the loss value of the BP neural network (BPNN) is about 0.28. The prediction accuracy of the WCGAN model can reach 86.3% when the training set is 5000, which is much higher than that of BPNN (77.8%). The reason is that the WCGAN model increases the number of generators according to GAN, which improves the low accuracy caused by pattern collapse. The IoT-based ITS can implement corresponding strategies according to the prediction results and control the access rights of nodes, thus ensuring the security of information resources effectively. The research content reduces the communication delay under ensuring the integrity and confidentiality of information in the process of data transmission, and provides a reference for ensuring the safe communication of IoV.