Aiming at the problem of speed jitter caused by high-frequency switching value in the sliding mode function in the speed regulation process of permanent magnet synchronous motor(PMSM) sensorless system of traditional sliding mode control an improved strategy using fuzzy super-twisting second-order sliding mode observer(FST-SMO) was proposed. The new observer is based on the high-order sliding mode control theory, which transfers the switching value of the sliding mode function to the high-order derivative, and uses the integral operation of the algorithm to weaken the chattering of the switching value. At the same time, considering that the gain of the traditional second-order sliding mode observer is fixed and cannot be adjusted according to the error, a gain self-adjusting method based on fuzzy control was proposed. According to the error value and its change trend, the current sliding mode gain was output by fuzzy algorithm, which enhanced the stability and robustness of the system. The simulation model of permanent magnet synchronous motor based on fuzzy supertwisting sliding mode system was built in Simulink environment. The speed regulation simulation was carried out in different speed regulation intervals, and compared with the speed fluctuation and rotor angle following of traditional sliding mode system simulation. The simulation results verify that the fuzzy super-twisting sliding mode observer has higher control accuracy and stronger adaptability, the speed is stable and the angle following performance is superior, and the chattering of speed has been reduced effectively.