The injection molding die configuration especially has a large influence on the product though the factor (resin temperature and Tamotsu pressure time, etc.) that influences exists in the product shape in the injection molding. It is produced by using CAE technology to need a lot of time and the cost for the die making now. In the present study, the metal mold design was optimized by paying attention to the warp that was one of the defective products, and using genetic algorithm (GA) that it was resin flow analysis and was the optimization technique. The analytical object is the injection mold die with several gates. The amount of the warp of this target product was assumed to be a displacement magnitude by using the amount of the warp as an index of the analytical result in the present study. To improve the total calculation time, the neural network is introduced in this optimization process instead of fluid analysis. If the neural network have enough teaching signal and can estimate the numerical result, it changes from the fluid analysis to the neural network on the way in the GA process. The remarkable improved results of optimization process for a calculation time is realized by using GA and neural network process as proposed in this study. The metal mold design was optimized by paying attention to the warp that was one of the defective products, and using genetic algorithm (GA) that it was resin flow analysis and was the optimization technique.