Abstract: With respect to the high similitude in spectrum of minerals' subclasses, it will be a challenge for applying visible and near-infrared method to extract the phyllo-silicate minerals. Take the Liuyuan in Gansu province as an example, through corresponding with the ASU Spectral Library, the spectral curve features of the images' endmember have been analyzed, the phyllosilicate alternated minerals (like Serpentine and Chlorite) have been identified on the basis of the curve features in spectrum of emissivity, and the mapping of alteration has been completed successfully. The results of field study show that, this method demonstrates high accuracy in terms of extraction and quantifies the effect of remote sensing in a more efficient way, which has an important significant for the mapping works of geological remote sensing.