Abstract: To take the capacity variations of road segments into consideration, and to avoid the fake traffic anomaly detections caused by the normal fluctuation of road segment capacity, a new traffic anomaly detection method based on travel time of path is proposed. Shenzhen road network as sub spatial regions is divided into grids, and then ST-matching algorithm is applied to match Shenzhen taxi GPS coordinate to corresponding road segments, after that the DBSCAN algorithm is adopted to compute the time-varying threshold of path travel time, the travel time above threshold is recognized as anomaly. While maintaining high accuracy and sensitivity of detecting traffic network anomaly, this method is cost efficient and easy to implement.