Crop health information represented through hyperspectral data is of great importance for precision agriculture. Because of the similarity in the spectral signatures of crops, discrimination of crop health using non-imaging spectral signatures is still a very challenging task for researchers. In this research work, spectral signatures are developed for soil, cotton, and maize crops from study area. Crop health is analyzed by considering of soil parameters and discriminated against Cotton and Maize crops. These all objectives prove to be essential in precision agriculture. ASD Field Spec4 for spectral signature collection has used, which has the capacity to discriminate objects in the range of 350-2500 nm. The study has carried out on various wavelength ranges or values. We have applied NDVI and CRI2 spectral vegetation indices for the analysis of spectral signature crops. Soil spectral signatures have been created and observed the N(Nitrogen), P(Phosphorus), K(potassium) and pH value of soil. Effects of various indices are studied and developed threshold values for health analysis of crops and found the relationship between soil health and crop health. Through the investigational study of results we found that NDVI and CRI2 performs well for crop health analysis. Finally Supervised machine learning algorithms SVM and KNN applied for classification of healthy and unhealthy crops in which SVM gives the result for health analysis of Maize is 90% and 87.5 for Cotton. KNN gives the accuracy of 85% for Maize and 92.5 for Cotton.