Abstract: In order to seek the better match between ECU (electronic control unit) and engine, the calibration test bench of HC4132UPS was developed based on MPC555. The sampling data was obtained by orthogonal experiment in test bench calibration. Using BP neural network, the mathematical model between control parameters and steady-state performance was built. The linear regression between the control parameters and power, fuel consumption and emissions was processed. The multiple correlation coefficient of output response was large than 0.94. The results showed that the network had good generalization ability and forecast performance. Using the neural network mathematical model as the constraints and objective function of performance optimization，the calibration was optimized by genetic algorithm. The experiment results show that this system can complete the collection of calibration data, and the calibration method based on neural network model and genetic algorithm is efficient and feasible.