Addressing the significant overestimation issue of traditional large-well method predictions for mine water inflow under multi-seam mining conditions, the Lvjiatao Mining Area was taken as the research subject. Leveraging the Surfer software platform, it integrated data on coal seam floor elevation, mining height, and geological structures to quantitatively identify the controlling coal seams for the direct water-filling aquifers (Layers III, IV, and ⅤA). Two key parameters were optimized:①For low-permeability aquifers interbedded with aquitards, the thickness of the aquifer exposed within the mining-induced fracture zone was used instead of the traditional full aquifer thickness.②For the exposed aquifer areas of controlling coal seams within the same aquifer, a spatial union operation was performed to eliminate errors from spatial overlap. Predictions using the optimized large-well method yielded a total mine water inflow of 9.54 m3/min. This showed a mere 2.9% relative error compared to the measured value of 9.27 m3/min, significantly outperforming the traditional method (which had an error >40%) and validating the reliability of the proposed approach.