Abstract: Using vector data as the initial location of the model, the paper presents a new active contour model which is used to extracting area water body from remote sensing images. In order to improve constringency speed and noise immunity, the image gravitation potential energy based on object-background gray value is added to the new model besides the traditional energy. In order to avoid the noise on the curve point of attraction and disturbance, the shape restriction potential energy based on discrete curvature is also utilized in the new model. The correlative parameters are obtained adaptively by taking full use of prior information of vector data. The paper also establishes the precision evaluating model based on similarity measurement, and introduces the process of solving based on greedy algorithm. The experiment demonstrates the superiority of the method.