Abstract: Due to the blindness searching in the final period,the Artificial Fish Swarm Algorithm(AFSA) can hardly obtain precision optima with high speed,according to the phenomenon of the division of labor,cooperation and competition that commonly found in society,this paper proposes a strategy characterized by the division of labor，cooperation and competition,which is designed by integrating several artificial fish with different algorithms. And among the fish,they keep the relationship of division,cooperation and competition,in which they achieve complementary advantages on improving the convergence property and accuracy. The test of application to some typical multimodal functions shows that this strategy has a better ability to locate the global solutions more efficiently and precisely，and to avoid falling into the locally optimal solutions. This strategy can also further improve the quality of solution by taking the adaptive strategy.

Key words: artificial fish swarm algorithm, division of labor, cooperation, competition