Abstract: Taking Beiliu River Basin in the Taihang Mountains as the research object, multiple linear regression and random forest regression were used to construct a soil organic matter inversion model to invert and estimate the surface soil organic matter content, in order to obtain the quantitative inversion and corresponding technical parameters of the surface soil organic matter in the Taihang Mountains.The results showed that, 1) methods such as Savitzky-Golay（SG） smoothing and wavelet packet analysis could improve the correlation between soil spectra and soil organic matter content.2) In the research on inversion of soil organic matter content, the random forest model had better model accuracy and effect than the multiple linear regression model.The random forest model constructed by using the low-frequency components obtained by the wavelet packet decomposition had the best effect in the inversion of SOM content, and R~2 was 0.812,achieving an effective estimation of soil organic matter.3)There were many factors that affect the accuracy of the soil organic matter content inversion model.In the future research,in-depth research on the quality of spectral data sources,spectral data processing methods and modeling methods was needed to further complete the soil organic matter content inversion technology system and improve the inversion accuracy of soil organic matter content.