In order to realize green manufacturing of Electrical Discharge Machining(EDM), energy consumption and pollutant emission are reduced as much as possible on the basis of ensuring machining efficiency and quality. Orthogonal experiment and nondominated sorting genetic algorithm (NSGA-II) were used to optimize the machining parameters. The parameters of current（I）, cycle rate （T） and efficiency（η） were selected as dependent variables, surface roughness（Ra）, energy consumption (EEV) and environment contamination (EEC) as evaluating indicator, and the EDM experiments of SKD11 were carried out. The correctness of the model between process parameters and response was verified by regression analysis, and the main factors affecting energy consumption and pollutant emission were obtained by signaltonoise ratio analysis. Finally, the regression relationship between the processing parameters and the machining effect was obtained.The NSGAII algorithm was used to optimize the parameters and got the Pareto frontier. The average relative errors of Ra, EEV and EEC prediction results are 6.46%, 10.45% and 9.58%, respectively, which shows that the optimization results are accurate and effective. Therefore, there is a certain guiding significance for research and green processing of enterprises.