The work aims to improve the accuracy and stability of the temperature control of the injection molding machine barrel, and solve such problems as nonlinearity, strong coupling and time variation in the multi-stage temperature control system of the injection molding machine. By analyzing the structure and working principle of the injection molding machine, the intercoupled circuits were decoupled as a single-circuit system with the static decoupling strategy based on the neural network; and the quantization factor of fuzzy PID was optimized by particle swarm optimization. The simulation and test showed that, the temperature response of each barrel was relatively fast, the barrel temperature overshoot was small, and the homeostatic control error was small, so that good control effects could be achieved. The simulation shows that, the proposed control strategy can overcome the coupling effects occurring between sections of the barrel. With strong adaptation, good stability and certain capacity of resisting disturbance, the temperature control effects of the injection modeling machine can be significantly improved, which has certain practical value.