Permanent Magnet Syschronous Machine (PMSM) has low torque, so good control is needed to be stable quickly. PMSM uses the principle of faraday experiments which is rotating a magnet in a coil or vice versa. When a magnet moves in a coil, there is a change in magnetic flux in the coil and penetrates perpendicular to the coil so that there is a potential difference between the ends of the coil, regarding this due to changes in magnetic flux. Magnetic flux can be changed by moving a magnet in a coil or vice versa by utilizing other energy sources. To get a good optimization result, the right control constants are needed. So that the best output is obtained. To get the right constants, a suitable and good method is needed, including using artificial intelligence. In this study using the Imperialist Competitive Algorithm (ICA) method. From the simulation results it was found that the best controller design in this study was ANFIS-ICA with the best profile, torque profile, voltage profile, and rotation profile. The largest current is 2.45 A, the smallest overshot torque is 0.48 pu, the largest voltage frequency is 9.84 khz, and the best rotation (close to the reference) is 700.02 rpm. The results of this study will be continued with the use of other artificial intelligence.