Abstract: In this paper,Fourier thermal infrared spectrograph was used to detect the soil spectroscopic data and organic content in arid region.Analyzing the characteristics of the original spectral emissivity,we obtained the first and second derivative of the original emissivity,built the models between organic matter and the original emissivity,the first derivative,second derivative respectively.The applicability of the models were examined and evaluated.The results showed that:① the soil organic content differed apparently,because of the different surface feature types and surroundings.②The change trend of thermal infrared spectral emissivity with the increase of the wavelength stayed the same.Soil thermal infrared spectral emissivity within 8~11.5 um was most sensitive to the organic content.③The emissivity of the first and second derivative was superior to the original emissivity apparently.Differentiating of emissivity can effectively enhance the correlation with organic matter and the second derivative of emissivity had the highest correlation with organic matter in the study area.④ The fitting functions of organic content with the first and second derivative of the thermal infrared spectral emissivity fitted well,especially the second derivative.The multivariate stepwise regression fitting model of the first derivative had slightly stronger predictive ability than linear function model.