Video watermarking algorithm hides meaningful watermarking information in a video sequence to protect copyright andauthenticate content. In this paper, firstly, we divide a video into several scenes by the histogram difference method of video frames,and extract quickly and effectively the key-frames of the video according to the statistical difference of the histogram difference ofvideo frames. Then, we propose a watermark embedding location selection strategy based on the combination of minimum distortioncost of image wavelet coefficient and Hash blocks encryption to determine the optimal watermark embedding target location. Final-ly, the watermarking is embedded into the DWT-SVD domain of the target location. The proposed algorithm obtains a good robust-ness, and achieves a better balance in the invisibility and robustness of the watermarking method.