The BP neural network control system is proposed on the basis of pressure flow control so as to control the speed and pressure during shield tunneling in complex strata. The physical model of shield thrusting system is established by AMESim; and then the BP neural network controller is designed by Simulink; finally, the system is simulated so as to analyze the response characteristics of thrusting speed and velocity of hydraulic thrusting control cylinder of shield machine under variable flow and variable load conditions. The simulation results show that: 1) Compared to conventional PID controller, the fluctuation amplitude of BPPID controller is lower and the adjusting response time is shorter. 2) The control accuracy of thrusting speed and pressure of shield machine under loading condition can be improved by using BP neural network PID controller. 3) The practice shows that the abovementioned controller has good stability and adaptability, so as to provide theoretical reference for design and optimization of shield control system.