Information about NPK content in farm soil is important for further application of the necessary dosage of fertilizer. At present, laboratory-based chemical analysis is being used to assess the status of soil nutrients, but these methods are complex, tedious, costly, and poor in in-situ condition. While, for precise farming, in-situ assessment of soil fertility status is of prime importance. Hyper-spectral remote sensing bears the potential of a detailed investigation of soil through analysis of its spectral absorption features. Therefore, this study aims to estimate quantitatively the content of fertilizer in a sample of soil making use of spectral properties of macro soil nutrients. For addressing the abovementioned issues, this study has used Derivative Analysis for Spectral Unmixing (DASU), approach consisting of spectral unmixing and spectral derivative analysis. The study reveals that the spectral region 993.2nm provides a unique feature. It leads to the development of a model for estimation of NPK fractional abundance in a soil sample. Further, this model has been validated for the good number of soil samples in a laboratory. Thus, the key contribution of this study is to underpin that; the hyper-spectral remote sensing may be used in-situ to estimate soil fertility of farm soil. Although, the fractional abundances of individual components of NPK may be considered as future scope of work.