Central heating system is a complex control system with the characteristics of time delay, nonlinearity and large inertia, and the effect of traditional PID control cannot achieve satisfactory results, and it also causes a certain waste of resources. Although the BP neural network PID controller improves the performance of the PID controller to a certain extent, the BP neural network itself still has some shortcomings. In order to improve the stability of the heating system and realize the rational use of heat, the particle swarm algorithm (PSO) is used to optimize the weights of the BP neural network PID controller. After designing the PSO?BP?PID controller, the simulation curve of the traditional PID control, BP?PID control and PSO?BP?PID controller is obtained by using MATLAB, and the improvement effect of system performance is obtained according to the comparison of curve effects.