In generally, the permanent magnet for optimal design is only used with a high level of residual flux density and energy product. In terms of cost and efficiency, it needs better ways to select suitable product of permanent magnet according to optimum design shape. Each optimal design of the synchronous motor obtained suitable product of permanent magnet for torque with ripple. It was selected the synchronous motor of 4.7kW power including bar and V type magnet. This paper is used a product of magnets by one group of NEOTEC series in TDK Company, and apply the FEM and the response surface methodology (RSM). As a result, the optimal region differs bar type and V type magnet depending on magnetic characteristic. It was obtained the optimal region considering torque with ripple and cogging torque with bar type and V type using RSM. It was found that the optimal region in bar type magnet which satisfies torque with ripple and cogging torque was created with 1.35(T) or higher residual flux density. On the contrary, the narrow optimal region in V type magnet was created with 1.28(T) to 1.34(T) residual flux density. And the optimal region in spoke type magnet was created with 1.27(T) to 1.35(T) residual flux density. It was obtained nine products in bar type magnet, two products in V type magnet, and three products in spoke type magnet considering the selected the optimal products of magnet using RSM. Of those products within the optimal region by each magnet shape, the one with higher energy product was selected as optimum product. In the selected 53B product in bar type magnet, torque is increased by 13%, torque ripple decreased by 4.4%, and cogging torque decreased by 34%. But, in the selected 42B product in V type magnet, torque is only increased by 5% which is based on the optimized directions similar to the reference product. In case of the selected 42SH product in spoke type magnet, torque with ripple and cogging torque is similar the reference product Based on those results, the optimal result was verified by correlation analysis and experiment for validity. The product has a correlation coefficient between torque and back EMF by 90% and has similar back EMF results between simulation with 54.56V and experiment with 54.7V. This paper proposed the optimal product of permanent magnet according to magnet shape and characteristic with same amount of magnet. And it was possible to select the product of magnet which had been performed the optimization using RSM.