The traditional video image coding method, that is to say, the intraframe coding based on DCT and the interframe prediction coding based on motion compensation is not suitable, for low bit rate compression and aside from this, the encoder is too complicated. The XYZ video image compression method based on 3-D DCT is discussed and realized we propose the 3-D Z scanning program of 3-D DCT coefficients, which was applied to the video image compression and has been proved to greatly increas the encoding efficiency.