Accurate and robust estimation of the vehicle sideslip angle is crucial for ensuring the safety and comfort of four-wheel distributed drive and brake electric vehicles. However, when onboard sensor fails and cannot provide reliable motion state information, it becomes necessary to isolate and substitute these sensor signals. To address these challenges, this paper introduces a fault-tolerant sideslip angle estimator that employs an unscented Kalman filter (UKF) for distributed drive and brake electric vehicles. Initially, a vehicle dynamics model is presented to facilitate vehicle motion estimation. Using onboard sensors, redundant signal analysis is conducted, and unreliable signals are detected and evaluated through a confidence level algorithm. Subsequently, adpative UKF is applied to calculate sideslip angle based on the reconstructed signal. The measurement covariance matrix is adjusted according to the confidence level to enhance the accuracy of the estimation. The introduced method is ultimately validated through Simulink-CarSim co-simulation. Simulation results indicate that, in the event of sensor failure, the sideslip angle estimation using the reconstructed signal exhibits only minor discrepancies compared to the actual values.