Abstract: The conventional methods of road change detection have disadvantages in terms of the data acquisition period, data cost, algorithmic complexity, calculation difficulty, and periodic updating. In this paper, by making full use of taxi GPS trajectory data distribution and timeliness, a new road network topology change detection method based on vehicle GPS spatio-temporal trajectories is proposed. In this method, the similarity between the GPS trajectory vector and the partial topological vector is measured using the vector similarity measure model, and the topological change of the road network is detected by comparing the path changes of the new, the waste and the reconstruction. Experimental results show that this method can not only detect the changes of new, wasted, and reconstructed parts of road network, but also can realize real-time change detection in urban road networks.