Salt content in the top layer of soil is one of main factors that should be taken into account in soil salinity classification.This study analyzed three saline soil reflectance spectral database which include soils with various salt and soil moisture contents.With the help of statistic analysis,the relationship between soil salt content and five hyperspectral indices was analyzed and contrasted.The optimal hyperspectral index and spectral predicting models for soil salt content was developed.The results showed that the most sensitive wavelengths for soil salt content were in the NIR and SWIR domains,indicating possible sensitivity of the characteristic features associated with salt minerals in the VNIR-SWIR region.The selected ND index can be calculated easily and is insensitive to soil moisture,and therefore has wide applicability for large-scale soil salt content mapping.