Abstract: The twelve months' radar data's backscattering coefficients was used to construct the time series for information extraction of rocky desertification in the year of 2016 in Salaxi and Huajiang demonstration areas,then ranked the karst rocky desertification information by fusing the SAR remote sensing image and optical remote sensing image,and verified the accuracy by field surveying.The result showed that the method of information extraction of rocky desertification based on multi-source remote sensing image data was more accurate and effective than traditional optical method in translating and ranking of rocky desertification.Their accuracy rates were improved by 7.69% and 5.99%,respectively,in mild and moderate rocky desertification region of karst plateau-mountain region,and their accuracy rates were improved by 9.41% and 5.39%,respectively,in moderate and severe rocky desertification region of karst plateau-mountain region.