Covered by vegetation results in the difficulty to monitor soil nutrition and its accumulation of farmland during crop's growing season by remote sensing, however, crop's growth status and growing speed directly reflect the soil fertility and fertilization, so it is a feasible method for monitoring soil nutrition and its accumulation of farmland based on remotely-sensed monitoring of crop's growth status. Two temporal NDVI were derived from hyper-spectral imageries collected at reviving stage and elongation stage, and two temporal grid-sampling data were interpolated to raster soil nutrition data using Ordinary Kriging analysis. The first soil sampling was processed before seeding, and another was after harvest time. It analyzed the correlation among remote sensing parameters, that is, NDVI(Normalized Different Vegetation Index), the absolute increment of NDVI around fertilization (⊿NDVI), the ratio of ⊿NDVI to NDVI obtained pre-fertilization(R_{_⊿} NDVI₁), and the ratio of ⊿NDVI to the summery of NDVI around fertilization(R_⊿ NDVI₂), and the increment of alkali-hydrolysable nitrogen in soil(⊿N) during winter wheat growing season. Results show that, compared with ⊿NDVI, NDVI obtained before fertilization can well estimate the increment of alkali-hydrolysable nitrogen in soil, and R_⊿ NDVI₁ is the best index for estimating nitrogen accumulation during crops growth period, while NDVI obtained after fertilization is the worst.