Abstract: The lifting scheme is a new bi-orthogonal wavelet constructing method, which can effectively reduce wavelet run time. The Embedded Zero-tree Wavelet (EZW) coding algorithm makes a full use of data distribution characteristics after wavelet transform, can realize progressive image transmission and is one of the most effective image coding methods. Yet, in the EZW algorithm, low-frequency and high-frequency data are coded by the same process, which would result in the loss of more important low-frequency data in the case of a high compression ratio. Significant loss of low-frequency data would lead to a poor recovery image quality. Furthermore, the EZW algorithm has problems of space complexity as well as temporal complexity. In view of those problems, this paper proposes three improvement methods. First, the low-frequency data are adjusted to the range of [0, 255] and then are saved directly. Second, a marking matrix is set up to track the coding position, instead of the dominant table and subsidiary table in the original process. Meanwhile, the dominant and subsidiary scans are carried out for each important coefficient simultaneously. Last, one more bit of precision to the important coefficients is added in the process of the subsidiary scan. The improved EZW and the integer lifting wavelet transform are combined and applied to a practical remote vision control system of a robot arm. The two results of image compression based separately on wavelet transform and DCT are compared and the proposed method shows better performance on both the recovery image quality and running time.