At present, many image copy-pastes tampering detection algorithms mainly compare the distance between feature points and threshold to complete feature matching, When the threshold is not selected properly, it is easy to cause some shortcomings such as error detection and omission detection. In this paper, an image copy-paste forgery detection algorithm based on multivariate describing factor combining correlation measure cluster is designed. Firstly, Forstner operator is introduced to extract the features of the image by calculating the Robert gradient of the pixels in the image. Then, the principal direction of the feature points is obtained by using the gradient feature of the pixels. Based on the principal direction, the multiple circular neighborhoods of feature points are constructed. The multivariate describing factor are constructed by using the gray level features, gradient features and Laplacian operators of the pixels in the neighborhood, and the feature vectors are formed to describe the feature points. Finally, the Euclidean distance function is used to match the feature points forward and backward to complete the feature matching. The correlation of feature points is measured by cross-correlation function, and image features are clustered to detect tampered content accurately. The simulation results show that the proposed method is not only more complete and correct than the current copy-paste tampering detection method, but also more robust than the current copy-paste tampering detection method.