<p>To achieve multi-fault tolerance control for vehicle suspension actuators, a fault tolerance strategy based on the Soft Actor-Critic (SAC) algorithm is proposed. This strategy adjusts output feedback controller parameters to modify the output force and incorporate additional compensation force. The output feedback controller is designed based on the vehicle suspension dynamics model, and the SAC agent is designed to adjust the output feedback controller parameters and achieve fault tolerance of the suspension system. A comprehensive main reward function and additional stage-specific reward functions are constructed to enhance vehicle ride comfort and constrain the dynamic deflection of the suspensions. In order to cooperatively control the ride comfort and these deflections under different driving conditions and failures, the weights of ride comfort and front and rear dynamic deflection in the reward function are adjusted in time by three-dimensional extension control to achieve the output adjustment of the vehicle suspension control system. Through MATLAB/Simulink simulations and vehicle tests, the SAC-output-feedback fault tolerance strategy is compared with normal and other fault-tolerance strategies. Results indicate that by the SAC-output-feedback fault-tolerant control strategy, the suspension performance is closer to that under normal conditions, the superior training efficiency and suspension performance are obtained with a gradient-based reward function, and the comprehensive fault-tolerant control performance of the suspension can be improved under different road surface grades and fault conditions.</p>