Focusing on the issue that the path calculated by traditional path planning algorithm is not necessarily the optimal path in reality, a path planning algorithm which combined the experience of taxi driving and took time as a measure was proposed. The implementation of this algorithm was to transform the path planning technology which took calculation as the center into data-driven mining technology which regarded data as the center. Firstly, the real manned trajectory data were extracted from a large number of taxi trajectory data and matched to the road network data. Then, the access frequency of the road segments were calculated according to the calculation results of map-matching, and Top- k road sections were selected as hot sections; Secondly, the similarity of road tracks between hot sections was calculated, and the trajectories were clustered to build k sections of hot road map based on the road network. Finally, an improved A ^* algorithm was used to calculate the optimal path. The experimental results show that compared with the traditional shortest path planning algorithm and the path planning algorithm based on hierarchical road network, the path planning method based on hot section map can shorten the length of the planning path and the travel time and improve the time efficiency of path planning.