Image watermarking technology has been developed rapidly, but the research of digital video watermarking is still difficult. In order to protect the copyright of digital video content, a novel video watermarking method using positive and negative quantization was presented in this paper. Some selected video frames were used to embed watermark information. The brightness component of the carrier video frame was divided into several blocks. Each block was implemented discrete cosine transform(DCT). The watermark was embedded in the low frequent DCT coefficients. The watermark embedding and extraction were based on positive and negative quantitative manner. In order to enhance the security of the algorithm, the watermark image and the random binary image were used as the inputs of the cellular neural network (CNN), and the output was used for embedding. The experimental results had shown the robustness of this method to common attacks. This method is simple, and is convenient for realtime video processing.