Multiprocessor task can be stated as finding a schedule for a general graph to execute on a multiprocessor system. In this paper an efficient harmony search algorithm (HSA) is proposed to solve the hybrid flow shop scheduling with multiprocessor task problems (HFSMTP). The best values of HFS's control parameters are determined by full factorial design. Computational results are compared with the genetic algorithm related to the HFSMTP at the literature. The result showed that the proposed HSA is effective for solving HFSMTP.