To improve the method for soil moisture monitoring, uncover the mechanism of moisture effects on soil spectral reflectance, and provide theoretical information for other soil properties’ spectral analysis and remote sensing monitoring, Black soil samples from Dehui city, Jilin province, China, their field and laboratory reflectance were taken, and derivate, continuum removal and statistical methods were used to analyze soil moisture spectral characteristics and build soil moisture spectral models. The results are as follows: there are five absorption vales on soil spectral reflectance in the region 400-2500 nm, the vales grow with increasing soil moisture; however, the absorption vales’ areas of crop residue in the spectral range of the first two vales of soil can be ignored, which can be used to determine whether soil is covered with residue; stubble field shows the characteristics of soil, the five spectral vales are significant, but for the effect of stubble, the minimum value of the second vale moves to longer wavelength, and the proportion of the right side decreases significantly, which indicates whether the field is furrowed or not; so soil remote sensing should be conducted after field furrowed; the correlation between soil moisture at the depth 0-5 cm and spectral data is more remarkable than that at the depth 5-10 cm; the relationship between soil moisture and reflectance is exponent or linear, which should be considered when other soil properties are remotely sensed; soil moisture spectral models are very stable and accurate, can be used for quickly soil moisture monitoring.