The aim of this work was to study the influence of extruder temperature on the quality of the cast film produced. According to the properties of the extruder temperature control system such as the time-varying, nonlinear, and the large delay characteristics, taking the single screw extruder as the object, we put forward a kind of partition ratio control method and designed the partition ratio type PID neuron network controller based on the strong decoupling ability of PID neural network and the characteristics of the temperature distribution of the material inside the extruder cylinder. Finally, the simulation and experiment were carried out to verify the method. The step signal overshoot of partition ratio type PID neural network controller was 5.9%, the pulse interference stability time was 162.8 s, and the power consumption during normal production was 12.04 kWh. In comparison, the step signal overshoot of traditional PID controller was 16.4% , the pulse interference stability time was 192.4 s, and the power consumption during normal production was 13.42 kWh. In conclusion, the robustness and overshoot of the partition ratio type PID neural network controller are well controlled, the control precision of the system has been greatly improved, and the energy consumption of per unit output has been effectively reduced.