New methods of optimizing the design of machines with high stiffness have attracted much attention. Conventionally, machine designers have carried out optimization by attempting to minimize static deformation or maximize static stiffness. Nevertheless, the dynamic behavior of the machine structure plays a deterministic role in the final machining precision. Therefore, we propose in this study an integral-stiffness-based optimization method for designing the optimal structure of a computer numerically controlled (CNC) grinding machine. The proposed novel optimization methodology includes a prototype designed on the basis of know-how and the determination of control parameters based on the mode shape, Taguchi’s experimental method based on finite element analysis (FEA), and grey relational analysis (GRA). The target parameters in the optimization are static stiffness, first natural frequency, and dynamic stiffness. Results reveal that the optimal structure of a CNC grinding machine obtained by merely considering the static stiffness exhibits good performance when applying static forces but inferior performance when applying dynamic forces. A good optimization approach for designing a high-precision machine should integrally consider the static stiffness as well as the dynamic stiffness. With our proposed methodology, machine designers can design an optimal high-stiffness structure of a CNC grinding machine more efficiently and accurately.