Abstract: The nonlinearity of the vehicle model and the uncertainty of road parameters seriously affect the robustness of the intelligent vehicle controller under certain working conditions, causing the intelligent vehicle to deviate from the planned path during driving. According to the vehicle two-degree-of-freedom dynamic model and the lateral error model, this paper establishes the lateral motion state equation, designs the lateral sliding mode control algorithm and uses the neural network to further optimize the control error of the system and external disturbances, with an aim of forming a lateral sliding mode variable structure control strategy. The stability of the control strategy is verified through the theory of Lyapunov stability. The Matlab/Carsim co-simulation model is established to simulate and verify the lateral control strategy of the sliding mode variable structure. The simulation results show that, compared with the single sliding mode control strategy, the maximum lateral control error of the sliding mode variable structure control strategy can be reduced by 38% under the condition of high speed and small turning angle, which shows a good robustness of the control strategy to the change of loads.