Abstract: According to the comparability between the Video Watermarking System and the Digital Communication System， an adaptive video watermarking algorithm based on LDPC is proposed in this paper. The watermarking information that is stochastically permuted and coded by LDPC is embedded in the modified DC components which are attained by Integer Transform twice. To balance between invisibility and robustness of watermark， it utilizes the length of the watermark and the value of the transformed coefficients to adaptively select the groups to embed watermark and the intensity of the coefficients modification. Experimental results show that the proposed approach can achieve fine video quality and can run faster than the algorithm based on DCT. Moreover， the watermark can be extracted blindly and it is robust to common video attacking. Especially， the robustness of the algorithm becomes stronger along with the decrease of the rate as the code rate is above the lower limit conducted by the Hamming Bound.