Taking the minimum pollutant emission and active loss as objective functions, this paper builds a multi-objective optimization model for power flow optimization of power system. In the proposed algorithm, an external archive of non-dominated solutions is kept which is updated at each iteration. Moreover, a method based on fuzzy set theory is employed to extract one of the Pareto-optimal solutions set as the best compromise one to provide the scientific decision basis for decision-makers. Simulation of IEEE-30 bus system and IEEE-57 bus system testify that this algorithm can avoid the local convergence effectively compared with other multi-objective optimization algorithm.