In order to realize the rapid H.265 codec method, a new H.265 codec method based on Compressed Sensing Theory is proposed. The H.265 codec method are divided into three different types, intra-frame type, interframe type and mix frame type. Combined with the Compressed Sensing Theory, the input images linearly transform into the residual data, scale and quantify the transformed coefficients; after loop filtering and adaptive compensation, the H.265 coding data are got. When decoding, the process is reversed. Experimental results show that this method improves the image video codec quality and the image video codec speeds.