The grey-brown alluvial soil in northern china was selected as research object, and the feasibility and possibility of real -time analyzing soil parameter with NIR spectroscopic techniques were explored. One hundred fifty samples were collected from a winter wheat farm. NIR absorbance spectra were rapidly measured under their original conditions by a Nicolet Antaris FT-NIR analyzer. Three soil parameters, namely soil moisture, SOM (soil organic mater) and TN (total nitrogen) content, were analyzed. For soil moisture content, a linear regression model was available, using 1 920 nm wavelength with correlation coefficient of 0.937, so that the results obtained could be directly used to real-time evaluate soil moisture. SOM content and TN content were estimated with a multiple linear regression model, 1 870 and 1 378 nm wavelengths were selected in the SOM estimate model, and 2 262 and 1 888 nm wavelengths were selected in the TN estimate model. The results showed that soil SOM and TN contents can be evaluated by using NIR absorbance spectra of soil samples.