Forests in mountainous regions play a critical role in global carbon regulation, yet biomass and carbon estimates for many dominant species, including Quercus ilex, remain limited in Pakistan's Hindu Kush Mountains. This study addresses this knowledge gap by quantifying the biomass and carbon sequestration potential of Q. ilex forests using an integrated framework that combines field-based allometric equations with machine learning techniques. A total of 60 systematically distributed sample plots were surveyed across three altitudinal zones, where diameter at breast height (DBH), tree height, canopy diameter, slope, aspect, and wood density were recorded. Aboveground and belowground biomass was estimated using two allometric models alongside Random Forest and Gradient Boosting Regressor algorithms to enhance predictive performance. The integrated modeling approach improved biomass prediction accuracy, with Gradient Boosting showing the best performance (R2 = 0.9925). North-facing slopes stored higher biomass due to favorable microclimatic conditions, while south-facing slopes exhibited reduced values under greater radiation and moisture stress. The mean carbon stock of Q. ilex forests was estimated at approximately 87.9 tonne ha-1, demonstrating substantial sequestration potential in semi-arid montane ecosystems. Overall, the study provides a novel and robust assessment framework that strengthens biomass estimation accuracy and offers valuable insights for sustainable forest management and climate-change mitigation initiatives in the Hindu Kush region.