Abstract: In order to address the problem of low electric efficiency of clutch asynchronous motors used abundantly in the industrial sewing machine control systems, an energy saving motor control system based on the DSP was designed. The industrial sewing machine control system is hysteretic, time-variant and nonlinear, hence the traditional PID control algorithm will reduce the serviceability and reliability of the system because the PID parameters cannot be adjusted timely when the controlled objects are varying. Considering this phenomenon, a late model named BP Neural Algorithm is added to the traditional PID control algorithm, whose self-learning and self-adjusting capabilities help realize the thoughts that parameters of a PID controller can regulate themselves as the system is fluctuating. Furthermore, applying the BP PID algorithm, MATLAB emulations of the energy saving motor control system are conducted, and the results of the emulations demonstrate that the control performance of BP PID is more superb than that of traditional PID. Lastly, the BP algorithm is synthesised in the software design, and achieves a satisfactory consequence.