The work aims to research the method to improve the control precision of dynamic quantitative weighing packaging system, with respect to its shortcomings, such as great inertia, hysteresis, non-linearity and inability to establish accurate mathematical model. An improved BP neural network PID of quantitative weighing packaging control system was proposed. By combining BP neural network and PID control method, and adjusting the self-learning and weighting coefficient of neural network, the parameters (Ki, Kp and Kd) of PID controller were optimized and the particle swarm algorithm was introduced into the neural network as its learning algorithm, so as to effectively improve the convergence rate of BP neural network algorithm. The simulation and experimental results showed that, the improved BP neural network PID control was featured by fast response speed, small overshoot and greatly reduced system weighing errors. The proposed control method can obviously improve the stability, accuracy and robustness of the quantitative weighing control process.