This article proposes a fractional order complementary non-singular terminal sliding mode control method based on neural network for permanent magnet synchronous motor (PMSM) control systems in electric vehicles. The method addresses the sensitivity issues caused by parameter variations, external disturbances, and friction. A mathematical model of the PMSM with uncertain factors is established, and a sliding mode controller is designed by combining generalized and complementary sliding mode surfaces to shorten the time for the state trajectory to reach the sliding mode surface and reduce chattering. A fractional calculus operator with filtering characteristics is used to improve position tracking accuracy and reduce steady-state error. A neural network is employed to estimate system uncertainties and compensate online, enhancing dynamic response and anti-interference abilities. Simulation results confirm the effectiveness and feasibility of the method, demonstrating its ability to improve control accuracy and support the development of electric vehicles. The PMSM mathematical model is derived based on field-oriented control, considering parameters like stator voltage, current, inductance, resistance, angular velocity, and flux linkage. The electromagnetic torque and motor motion equation are established, with the speed tracking error defined as the difference between the reference speed and actual speed. The model is adjusted to account for parameter variations and uncertainties. The fractional order complementary sliding mode controller is designed by combining generalized and complementary sliding mode surfaces, using fractional calculus to enhance performance. A neural network is integrated to estimate and compensate for system uncertainties, improving dynamic response and anti-interference capabilities. The method shows fast response, high tracking accuracy, and small steady-state error, effectively suppressing external disturbances. The paper concludes with the organization of the content, mathematical model, controller design, simulation results, and conclusions.This article proposes a fractional order complementary non-singular terminal sliding mode control method based on neural network for permanent magnet synchronous motor (PMSM) control systems in electric vehicles. The method addresses the sensitivity issues caused by parameter variations, external disturbances, and friction. A mathematical model of the PMSM with uncertain factors is established, and a sliding mode controller is designed by combining generalized and complementary sliding mode surfaces to shorten the time for the state trajectory to reach the sliding mode surface and reduce chattering. A fractional calculus operator with filtering characteristics is used to improve position tracking accuracy and reduce steady-state error. A neural network is employed to estimate system uncertainties and compensate online, enhancing dynamic response and anti-interference abilities. Simulation results confirm the effectiveness and feasibility of the method, demonstrating its ability to improve control accuracy and support the development of electric vehicles. The PMSM mathematical model is derived based on field-oriented control, considering parameters like stator voltage, current, inductance, resistance, angular velocity, and flux linkage. The electromagnetic torque and motor motion equation are established, with the speed tracking error defined as the difference between the reference speed and actual speed. The model is adjusted to account for parameter variations and uncertainties. The fractional order complementary sliding mode controller is designed by combining generalized and complementary sliding mode surfaces, using fractional calculus to enhance performance. A neural network is integrated to estimate and compensate for system uncertainties, improving dynamic response and anti-interference capabilities. The method shows fast response, high tracking accuracy, and small steady-state error, effectively suppressing external disturbances. The paper concludes with the organization of the content, mathematical model, controller design, simulation results, and conclusions.