This study proposes a novel identification method of critical roads based on the combination of GPS trajectory data and directed weighted complex network. First, with both the static road network topology and the dynamic traffic flow characteristics taken into account simultaneously, a new spatial temporal model of urban transportation, namely a directed weighted complex network, is proposed. Then, combining the structure of road network with the strength influence of traffic between adjacent roads, a mixed influence-based identification algorithm of critical roads is proposed. Finally, we analyze taxi-GPS trajectory data collected in Lanzhou, China. We perform a comprehensive analysis to visualize the spatial–temporal changes of taxi services, critical roads, and critical intersections. Moreover, the correlation coefficient has been used to evaluate the performance of the identification algorithm of critical roads. The results show that the new identification algorithm is more effective and practical than traditional congestion index analysis. Our investigation should be helpful in urban traffic management and the residential choice of alternative routes.