In this paper, we propose a design method of fractional order PID controller via hybrid of electromagnetism-like (EM) algorithm and genetic algorithm (GA). The hybrid algorithm improves the performance of EM algorithm by using GA, called IEMGA. The main modification is that the randomly neghiborhood local search is changed by genetic algorithm for designing fractional order PID (FOPID) controllers. IEMGA combines the advantages of multiple search, global optimization, and faster convergence.It does not need any gradient information and it is capable of decreasing the computational complexity. Simulation results show that IEMGA has the ability of golbal optimization and faster convergence. In addition, it improve the computation complecity of EM.