Neural network is used to solve the nonlinear problem, and predictive control is used to overcome the dead-time problem. Aiming at expansion valve control loop in refrigeration system, the predictive control based on neural network is proposed to control the superheat in evaporator. The simulation and application show that this method is feasible with good dynamic responses and stronger robustness, and provide an efficient control of superheat in evaporator. The control results are better than that of PID control.