To overcome difficulties of the traditional ant colony optimization, a novel ant colony system based on dynamic search(DSACS) strategy for path planning problem of mobile robot is proposed. Therefore, a dynamic search model is designed. in the prophase, a bigger parameter is used to increase the diversity of the population; in the anaphase, a smaller parameter is adjusted through the attenuation model to accelerate convergence. Experimental results of TSP benchmark instances show that the improved ant colony algorithm can not only accelerate the convergence, but also improve the quality of the optimal solution. Simulation results of path planning problems under the complex environment verify the cutual effect of the DSACS strategy.