A new compression scheme of satellite interferential superspectral images is proposed based on the analyses of satellite superspectral images. There are two novelties in the scheme: the first is that the compression coding is an image sequence compression method with low memory, which utilizes the matching algorithm in the wavelet domain to remove the redundancy between the two successive frames; The second is that in order to protect the information of the spectrum the scheme adopts a new ROI (Region Of Interest ) coding algorithm, in which the lifting procedure is carried out in code stream organization stage. As results, the low coding efficiency situation due to the nature of the bitplane shift methods could be improved, which enables the high fidelity in the background (BG) and makes better qualities of spectrum curve. The experimental results