In view of the shortcomings of ant colony algorithm, an improvement method has thus been proposed aiming at an improvement of the convergence speed of the algorithm and reducing the number of path turning points. Firstly, with a corner heuristic function is designed, the distance heuristic function is improved for an enhancement of the directionality of the target node towards the path finding ants, with unnecessary corner turns in the path reduced. Adding adaptive dynamic variables on the basis of probability transfer formula helps to increase the search space in the early stage of the algorithm, thus improving its convergence speed in the later stage. Finally, partition rewards and punishments are applied to different ants so as to expand the inspiration of pheromones. The experiment shows that under the two different complexity environments, compared with traditional ant colony algorithms, the number of path turns is reduced by 58.8% and 38.1%, respectively, with the efficiency improved by 90%. Moreover, the computational efficiency is not easily affected by environmental complexity, thus verifying the superiority of the algorithm.