The paper aims to improve the precision of bag length and the appearance quality of bag. A new speed control method for packaging material of three-servo pillow type packing machine based on neural network PID adaptive control was proposed. The traditional PID control method was combined with the neural network control. A neural network PID controller was designed, including its structure and learning algorithm, which can be used to solve related nonlinear problems. The simulation and experimental results showed that when the neural network PID control method was adopted and the speed of packaging material became stable, the time was about 2 s and the maximum overshoot was less than 2%. The bag length error can be effectively controlled within ±1 mm. Compared with the traditional PID control, the designed control method has the advantages of fast response speed, strong anti-interference ability, stable control output, etc. It can significantly improve the length accuracy of the packaging bag.