A coding approach based on modified embedded zero-tree scheme is proposed for image compression. At first, an image is decomposed using lifting wavelet transform. Then, the obtained wavelet coefficients are encoded by the proposed modified embedded zero-tree scheme. Finally, adaptive arithmetic coding is adopted to encode the coding coefficients. Lifting wavelet transform accomplishes in-place computation, fast forward/backward transformation, and integer-to-integer transformation, as well as modified embedded zero-tree scheme takes the statistical distribution of image data to encode wavelet coefficients; thus, an optimal coding approach is obtained for lossless data compression and progressive transmission. Experiments show that proposed approach achieves the best compression ratio compared with JPEG, EZW, and SPIHT schemes.