In order to improve the robustness of image Hash algorithm a new image Hash algorithm based on discrete curvelet transform is proposed. The image is firstly preprocessed

and then decomposed with discrete curvelet transform via wrapping. The curvelet coefficients of low frequency contained the main features of image and the coefficient of details of two layer contained rich edge information are as the feature vectors. And Logistic equation is used to encrypt the eigenvector. Finally

the image Hash sequence is obtained by quantization and compression. Experimental results show that the algorithm has better robustness compared to some other Hash method. It is fragility to different images. The chaos system enhances the security.