Aiming at the problem of poor accuracy of feature points matching in images with large deformation，an image matching method based on weighted shape context under integer constraints is proposed.As graph theory is used to describe the structure information between image characteristics，feature points matching is converted to a problem of graph matching.Shape context is introduced based on the construction of similarity measuring.Possible matching points between graphs are used as nodes of an assignment graph，and the statistical characteristics of weighted shape context within graph point sets and between graph point sets are used as similarity measure function to construct weights of graph edge.At last，the graph matching problem is resolved by iterative solving method under the integer constraints.Experiments on real image feature points matching demonstrate that the proposed method is effective.