Hyperspectral Imaging (HSI) is a method that uses data collected from across the electromagnetic spectrum range to calculate the gamut of individual picture pixels. By studying their distinctive spectral fingerprints, objects and materials may be identified. HSI was originally developed for military and applications in space, but they are currently found in climatology, agriculture, land use/cover analysis, and Earth Observations through Remote Sensing (RS) satellite images. The EO-1-H (Earth-Observing-One-Hyperion) HSI dataset was acquired from 2008 to 2015 from CONUS. This chapter employs the Extreme Gradient Boosting (XGBoost) model, well known for its high efficiency, classification, and regression problems. Additionally, it utilizes the method of Non-Negative Matrix Factorization (NMF) for data dimensionality reduction. The chapter covers an exploratory data analysis using diverse methodology explanations. Multiple Machine Learning (ML) classification models were compared to the suggested model, yielding better accuracy at 89% over training and test data.