With regard to the copy-move forgery of video frame sequence, this paper proposes a algorithm for video detection and recovery base on the correlation of the contents between frames. Firstly, the video frames are translated into a series of consecutive image frames, and by blocking the frame to extract eight dimensional feature vectors. With Euclidean distance computed the correlation between frames, a deviation matrix is import in the algorithm to reduce the false alarm rate. A threshold value is used to localize the modifications to a frame level and recover the original video. Experimental results show that this method can effectively detect the copy-move forgery of video frame sequence and video restoration.