<p>Forest fires play a pivotal role in altering soil properties, influencing forest ecosystem dynamics and regeneration potential. This study investigates the effects of low-intensity forest fires on the physicochemical characteristics and nutrient dynamics of soils in oak (<i>Quercus leucotrichophora</i>) and pine (<i>Pinus roxburghii</i>) forests in the Western Himalayan region of Uttarakhand, India. The primary objective was to compare the soil properties of burnt and unburnt (control) sites, focusing on macro- and micronutrients, and to assess the applicability of a k-nearest neighbor (kNN) machine learning algorithm for classifying soil samples. The results revealed that burnt sites exhibited significantly higher pH, electrical conductivity, cation exchange capacity, total nitrogen, available phosphorus, and exchangeable calcium and magnesium, while concentrations of micronutrients (Fe, Zn, Cu, and Mn) were lower due to fire-induced chemical transformations. Correlation analyses showed weaker interrelationships among soil variables in burnt sites, indicating altered nutrient cycling. The kNN classifier successfully distinguished burnt from unburnt soils with 97.4% accuracy, confirming the efficacy of using 17 soil parameters to characterize post-fire soil conditions. These findings underscore the profound influence of forest fires on soil quality and highlight the potential of machine learning in pre and post-fire assessment.</p>