The real-time compression of the remote sensing image is one of a problem in urgent need of satellite remote sensing technology, the development of the compression method suitable for real-time remote sensing image has an important theoretical and practical value. In response to this problem, a fast compression method which is fit for remote sensing image is designed based on the features of decomposed image and the statistical analysis of the sub-images. The method uses different processing methods for the every sub-image compression after lifting wavelet. The sub-image with low frequency is compressed lossless by DPCM method, and the improved EZW coding and RLC method was used to compress the sub-images with high frequency. Through extensive experiments, the results show that the new compression method is more effective than the others. In addition, the new compression method is simple, and the memory requirement in the operation process is lower, and it is fit for parallel optimizing progress.