This paper is concerned with the problem that the calculations of Dijkstra algorithm increase rapidly with the amount of data. First of all, the minimum binary heap as the auxiliary data structure of Dijkstra shortest path algorithm is proposed, which can effectively reduce the  calculations and improve the operation efficiency. Then, the distance data and train accessibility data of 31 provincial capitals in China (excluding Hong Kong, Macao and Taiwan) are chosen in the numerical simulations. Finally, the results indicate that it is much faster to find the shortest path by applying the binary heap optimization algorithm than the traditional algorithms, and it is more obvious when increasing scale of the problem. It is conclued that the binary heap optimization algorithm is able to significantly improve the computational efficiency.