The aim of this paper is to propose a global minimization algorithm belonging to the class of Controlled Random Search algorithms, to interface with the Finite Element analysis procedure for optimizing the design of Buried Permanent Magnet Synchronous motors and Synchronous Reluctance motors. The behavior of the proposed algorithm has been compared, in terms of attainable optimum and computation time, with the one of a minimization procedure available at commercial software. This comparison has shown a significant more efficiency of the considered algorithm especially in the more complex designing of a Synchronous Reluctance motor. The goodness of the results achieved with the proposed algorithm shows that combining the FE analysis with a suitable strategy is an effective and powerful tool for the design optimization of electric motors.