[Objective] The measured spectra data of several typical bare surfaces were analyzed such to provide references for desertification prevention and control. [Methods] Methods of spectral analysis as moving average, reciprocal take logarithm(RLR) reflectance, first derivative transformation(FDR) and continuum removal(CR) were used to identity the sensitive band range of each ground feature. The paper aimed to explore the application of GF-1 and other extended use of that. The spectral indexes, such as normalized difference index(NDSI), simple ratio index(SR), spectral difference entropy index(SDI) were used to separate each other. [Results] Differences of the original spectral signatures among several bare surfaces were small. Spectral reflectance transited slowly from invisible value to near-infrared one, but the rate of increasing was different among each ground feature. The differences among bare surfaces could be magnified by first order derivative transformation, reciprocal logarithm transformation and continuum removal spectra. The effectiveness of transformation was all better than original spectrum. [Conclusion] Because the calculated difference was magnified between the original and the transformed spectral index, the first order derivative transformed SDI can distinguish the surface features much better.