The issue of CNC machine tool energy consumption and environmental protection plays an important role on manufacturing technology research since CNC machine tool energy is consumed during motor racing or cutting process. The paper analyzes CNC machine tool energy consumption influence by cutting parameters of cutting speed, feed speed, cutting depth. Based on nonlinear mapping ability of neural network, the model of CNC machine tool energy consumption related to cutting parameters is established by using experimental data, and then the optimal combination of cutting parameters are searched by using global optimization of genetic algorithm, and verified in CNC machine tool cutting experiment. The proposed method provides a good energy control proposal for CNC machine tool roughing process. The experimental results show that the energy consumption is optimal. © (2012) Trans Tech Publications, Switzerland.