The minimum detection error threshold of packaging and printed image perceived by human eye was calculated according to contrast detection characteristics of human eye. The DCT transform coefficients of image were quantified according to threshold value, encoded with Huffman coding method, and verified with experiments. The results showed that PSNR got to 49.647 1 dB; compression ratio was 11.497 7; and coding efficiency was 0.743 8. The result indicated that performance parameters of objective evaluation for data compression achieved better effect, and difference was little between the image decoded and the original image. From subjective evaluation, the eye could hardly distinguish the differences, which indicated that the image compression technology used did not affect the quality of the image. Results also showed that the compression algorithm can not only achieve some compression ration, but also does not affect the observed effect of printed image, so that it can satisfy the requirements of packaging and printing image transfer.