In order to determine the main disturbance variables and internal coupling relationship in the control process of rubber compound extruder and to better realize the model identification adaptive control and precise decoupling control for temperature and pressure control system, by combining radial basis function (RBF) neural network and PID neuron structure, a model based on RBF neural network identification model and adaptive control was designed to complete the model identification and adaptive control for the melt temperature and head pressure, and the optimized RBF neural network was used for precise decoupling control. The identification model of the temperature and pressure coupling system was established by using MATLAB software, and the control effect was compared with the traditional identification model and decoupling method. The results show that the system based on optimized RBF neural network has better identification ability and can achieve adaptive decoupling control; the coupling identification model established by using optimized RBF neural network has ideal coupling identification and decoupling effect, which can improve the precision of temperature and pressure control system and the quality of extruded semifinished products in a certain extent, and realize precision extruding.