Abstract: For the characteristics of multi-objective optimization problems is proposed for multi-objective optimization problems ant colony algorithm.Definition of evolutionary algorithms selected from the population when a certain number of individual sources as the center of pheromone diffusion, more than the distance between the centers of intervals;group of other individuals in accordance with the distance from the nearest source of the principle of individual ownership in one of the Pheromone diffusion source;each spread source of pheromone pheromone diffusion algorithm in accordance with the individual to obtain from the center of the pheromone;each generation groups in the center point to the next generation population retained to ensure convergence and maintenance groups Diversity.Finally, multi-objective knapsack problem to test algorithm performance, and with the MOA and the NSGA-II algorithm was simulated compared.Results show that the search efficiency, the true Pareto front approximation to good effect, to obtain the spread of a wide range of solutions, is a multi-objective optimization problem solving and effective method.