Abstract: Often a classical GA is not flexible or effective for a optimizing application of complicated multimodal problems. It is difficult to reach more than one of the global maximum, simultaneously. Considering the above deficiency of GA, the paper presents a multimodal partheno-genetic algorithm (MPGA). The aim of the proposed algorithm is to find not only a single solution but also a set of the optimal. Both new a crossover operator and selection schemes are introduced so that every population individual inherits more genetic material of its own forefather. The diversity of population individual can be improved greatly in this approach. The size of population is very flexible and the architecture of computation is well suited for parallel implementation. The simulating cases show the effectiveness of this approach.