<p>The mechanical behavior of rock masses in longwall mining is critically influenced by spatial correlation among material properties, yet conventional deterministic models often overlook this variability. Conventional deterministic models often overlook this spatial variability, leading to potentially misleading assessments of rock strength and stability. This research addresses the critical need for a nuanced understanding of rock mass behavior by investigating the effects of spatial correlation on stress distribution and failure mechanisms in coal seams. The primary objective is to evaluate how incorporating spatially correlated random properties can enhance the accuracy of predictions in mining operations. The study uses a three-dimensional numerical model to contrast deterministic approaches with a stochastic framework that integrates spatial correlation factors. The methodology involves generating a realistic random field database based on the Extreme Value stochastic model, which is then applied to simulate stress responses in the rock mass under various loading conditions. This 40% reduction in peak stress estimates translates to substantially different safety assessments and mining strategy recommendations compared to traditional deterministic approaches. This research underscores the necessity of adopting stochastic approaches in rock mass evaluations, as they provide a more accurate representation of real-world conditions. The insights gained from this study are essential for developing safer and more effective longwall mining strategies, highlighting the importance of considering spatial variability in rock mechanics. The findings contribute to the advancement of mining engineering by integrating advanced statistical techniques with practical applications, ultimately enhancing operational efficiency and safety in mining practices.</p>