In order to improve the yaw stability of hub motor-driven electric vehicle under different road adhesion coefficient, the stability control system of hub motor-driven electric vehicle was designed. In order to improve the control effect of the control system, the unscented Kalman filter observer is built to observe the driving state and adhesion coefficient of the vehicle in real time. The control strategy of braking torque distribution based on Unscented Kalman filter state observation is proposed. The control strategy adopts hierarchical control structure, and the upper layer adopts fuzzy neural network control to calculate the desired yaw moment for stable driving; According to the observed adhesion coefficient, a single wheel / single side differential braking control strategy is proposed. The simulation results show that the yaw stability control strategy can make the vehicle"s yaw rate follow the ideal value better under the condition of high and low adhesion coefficient, and the sideslip angle deviation of the center of mass is controlled at 0.4° And 0.3°， In this way, the stability and safety of hub motor electric vehicle are improved.