Abstract: Ship track data are processed with sliding window algorithm to find the ships in anchored condition. The trajectory of each ship at anchor is processed with K-means clustering algorithm and the clustering center is taken as the anchoring position of a ship. The Voronoi regions are defined according to the position data set of the ships at a particular time and combined to form a Voronoi graph. The area for a ship to anchor in the anchorage is extracted from the size of related Voronoi regions. The method is verified with data from Zhenjiang section of the Yangtze River.