Pursuiting cutting efficiency and surface quality during machining monocrystalline silicon can improve the quality of the wafer itself and reduce the production cost. The second-order model of variable parameter voltage,pulse on time,pulse interval,wire saw speed and process result is established by response surface method,The variance analysis shows that the model equation is significant. The improved non-dominated genetic algorithm（ NSGA-II） is used to optimize processing process,and we can get a set of non-dominated solutions. These non-dominated solutions can satisfy the needs of engineering workers under different conditions. The corresponding experimental proving the non-dominated solution is feasible.