The traditional Dynamic Route Guidance System (DRGS) provides only the optimal path to the travelers, which may easily lead to aggregative response of the travelers and overcrowding drift. This paper presents an approach based on Ant Colony Optimization (ACO) for solving the k-shortest paths problem in DRGS. In order to improve the convergence rate, the basic ACO is improved by introducing direction function the weight coefficient of which can be adjusted to vary state transition rule and standardized transformation to eliminate the influence of the size and dimension of pheromone and heuristic information. Compared with basic ACO, simulation experiments indicate that the improved ACO is more effective and efficient.