The work aims to establish a mathematical model of film tension in the process of film mulching, in order to solve the problem of film tension instability in the production process of suspended packaging. Aiming at the shortcomings of traditional PID control in film tension control, an adaptive PID film tension control method was designed based on BP neural network. The parameters of PID controller were adjusted according to the running state of the system. Through the self-learning of neural network, the weight coefficient was adjusted to ensure the stable working state of controlled object. In order to further optimize the control performance of BP neural network, the initial threshold and weight optimization were realized with the fish swarm optimization algorithm. Finally, the simulation and experimental research were carried out. Simulation and experimental results showed that, the system’s response speed was fast and the maximum overshoot was small during tension control based on the improved BP-PID control algorithm. The maximum relative error of film tension was only 0.5 N, which was relatively small. The improved BP-PID algorithm has high control accuracy and stability, and can meet the requirements of packaging coating.