An orthogonal experiment for high-strength mould steel NAK80was designed with cutting speed, feed rate and cutting depth as relevant parameters. According to the experimental results, statistical models of cutting force and surface roughness were built up using a linear regression method. Superadded the theoretical formula of material removal rate, a multi-objective optimal model with three objective functions was built up and optimized by applying non-dominated sorting genetic algorithm-II (NSGA-II),to study the optimization of parameters in milling process for high-strength mould steel. Several cutting parameter combinations conforming to the requirements were acquired. The results show the proposed method can get the optimal cutting parameter combinations and guided the actual production.