In this paper, a systematic approach to achieve optimized design of interior permanent magnet machine (IPM) having novel semi-overlapping windings (NSWs) is presented. The optimization parameters have been determined individually by performing sensitivity analyses. Multi-objective global optimization is subsequently performed. Genetic algorithm (GA) approach, which is also known as “random search with learning algorithm” and an effective optimization tool used for design optimization of electric machines, is employed. IPMs equipped with integer-slot distributed winding (ISDW) and NSW are initially designed by using the geometric and operating parameters of Toyota Prius 2010 IPM. Subsequently, a time-stepping 2D finite element analysis (FEA)-based program is employed to perform the optimization and quickly evaluate the optimal solution among the thousands of design candidates thanks to the sensitivity analyses. In order to reveal the effectiveness and rapidity of the multi-objective global optimization, a comprehensive electromagnetic performance comparison between the original (with ISDWs), initial and optimal designs (with NSWs) is presented. Finally, a small IPM prototype globally optimized by using the proposed procedure is manufactured, and the FEA results have been validated by measurements. Our goals in this study is to advance the state of the art in the multi-objective design optimization of NSW IPMs by performing sensitivity analyses to reach the optimal solution quickly and to determine the most sensitive design parameters affecting the key performance characteristics.