The protection of intellectual property rights became recently a pressing need especially with the rapid growth of transmission techniques. In this paper, we present as a copyright protection method, a blind video watermarking technique based on video scene segmentation and 3D wavelet transform. First, a gray scale image as copyright sign is decomposed with different resolution and embedded in the corresponding resolution of the decomposed video. As the multiresolution representation of the video is for spatial axis of each frame as well as the temporal axis of the video, it is convenient for watermark to be adapted to motion and motionless regions of the video and the spatial detail of each frame. The experimental results show that this technique has more visually recognizable information of copyright ownership. In addition, these results proved its robustness against several video degradations and watermarking attacks including frame dropping, frame averaging and lossy compression.