In order to rationally select optimize cutting parameters in turning, the cutting parameter optimization method for low energy consumption and low surface roughness was proposed. The input power consumption and surface roughness were obtained by the orthogonal dry turning tests, in which 45 steel was carried out on CAK3665ni lathe. Based on the analysis of input power and surface roughness, multiobjective genetic algorithm was used to optimize the cutting parameters by a model, which regarded the specific energy consumption (SEC) and average roughness (Ra) as goals. Compared with empirical data, it is shown that optimized cutting parameters will reduce the SEC and Ra in cutting significantly.