Globally, grasslands constitute one of the largest land ecosystems that sequestrate substantial amounts of Soil Organic Carbon (SOC) and provide various ecological services such as regulating the climate, improving water quality, and supporting biodiversity. Woody encroachment has pronounced impacts on grassland ecosystems, exerting substantial influence on their structure, ecological dynamics, and SOC distribution. To the best of our knowledge, concerns and discrepancies about the influence of woody proliferation on SOC within protected grasslands remain largely underexplored. Remote sensing offers cost-effective, time-efficient, and environmentally friendly means to quantify SOC distribution and variability in grasslands. Consequently, this chapter sought to quantify the spatial variability of SOC across a pristine and woody encroached grassland using PlanetScope image data. Employing a stratified sampling technique, 254 samples were collected, and the Loss-on-Ignition procedure adopted to derive SOC. The PlanetScope spectral bands and derived vegetation indices were utilized to quantify SOC variability between woody encroached and pristine grassland. A tenfold cross validation technique was adopted to train the Deep Neural Network (DNN) model using 21 input variables including SOC values. To improve model accuracy, hyper-parameter tuning was employed. The findings indicate a higher concentration of SOC sequestration across landscapes affected by proliferation of woody plants compared to pristine grasslands. The DNN model produced acceptable accuracies with Root Mean Square Error of 1.90 t/ha and R2 value of 0.64. The research establishes a framework for continuous assessment of SOC stock in protected grasslands.