Estimating the content of soil organic matter (SOM) using hyperspectral remote sensing is an inevitable requirement for precision agriculture. In this study, averages of spectral reflectance and first derivative of 791 sample pieces were measured in the visible and nearinfrared (350~2500 nm) bands, which belong to five different soil types from four provinces, including phaeozem in Heilongjiang, alluvial soil in Henan, paddy soil in Sichuan and Zhejiang, purple soil in Sichuan and Zhejiang, and seashore saline soil in Zhejiang. The correlation coefficients between organic matter content and spectral reflectance were calculated, while the characteristic spectral bands mentioned in precedent researches were summarized. The results show that 600~800nm can be used as common characteristic bands of these different types of soils from different regions in the study, which is important for building the universal model for soil organic matter with alternative parameters. Another conclusion is that above 2% in SOM is not necessary for predicting SOM.