The shortest path finding problem is the core of route optimization algorithm. And the shortest path finding is one of the most important modules of the traffic navigation system. There are many kinds of route optimization algorithms, including some famous algorithms like: Dijkstra, Bellman-Ford, Floyd-Warshall, A* algorithm and so on. Even the same algorithm has a variety of different implementations. This paper provides an improved route optimization algorithm by improving and researching on the memory occupation and searching time consumption of Dijkstra algorithm. Finally, simulation results show that the improved algorithm has a higher efficiency than the original.