<p>In order to solve the problem that path following control algorithms for intelligent vehicles are difficult to improve following accuracy due to the deployment in vehicular controller with limited computational capability, this paper proposed a path following control system based on vehicle-cloud cooperation. Firstly, the signal transmission process of the system and the vehicle dynamics model considering tire nonlinearity and the steering system model were established. Secondly, a cloud nonlinear model predictive control (NMPC) is designed based on the directly discretized nonlinear vehicle dynamics model, and then combined with Unscented Kalman filtering (UKF) for vehicle state estimation and control output prediction of the cloud controller after the communication delay is calculated by the timestamp method. Thirdly, the vehicle tracking strategy and flexible fuzzy control strategy are designed to realize the cooperative path following control between vehicle and cloud while avoiding the jitter caused by direct controller switching. Finally, simulations and cloud-in-the-loop tests are carried out respectively, and the results show that the strategy of vehicle-cloud cooperative control significantly improve the tracking accuracy and stability under the nonlinear operating conditions of the vehicle.</p>