Abstract: Considering that the traditional ant colony algorithm converges slowly when solving the shortest path problem and easily falls into the local optimal solution, we propose an improved ant colony algorithm based on a heuristic mechanism. On the basis of the ant colony system (ACS) algorithm, the heuristic function is dynamically adjusted according to the distance between the candidate node and the target point to improve the convergence speed. When the algorithm falls into a local optimum, a penalty function is introduced so that the pheromone on the current optimal path decreases rapidly, and the effect of the positive feedback reduces at the ant's next search to prevent the algorithm from falling into a local optimum. Simulation experiments show that in a complex environment, including concave obstacles at the end point, the algorithm exhibits good performance in both the quality and convergence speed of the solution.