Because of the traditional ant colony algorithm based on positive feedback search way, lead to the existence of slow convergence speed and shortcoming of easily trapped in local minima. This paper propose a kind of extended ant colony algorithm based on mixed feedback mechanism(MF-ACO), the algorithm on the basis of traditional ant colony algorithm, define a kind of extension type ants which have strong global search ability, help algorithm out of local minima; In addition, the negative feedback balance mechanism based on stimulus-response model, the convergence ability and global search ability of dynamic balance algorithm were designed by referring to the labor division behavior of ant colony. Finally on the basis of division of labor model of individual ants pheromone update strategy was improved, to further accelerate the algorithm convergence speed Based on multiple TSP instance as the test object simulation experiment, the experimental results show that the superiority of the algorithm in this paper, and then use this algorithm for robot path planning problem, to verify the effectiveness of the proposed algorithm in actual application.