Abstract: To identify a more effective method for predicting water inflow in production mines, this study first examines the "large well" conceptual model. The formation of "large wells" in coal mines typically requires several decades, making it difficult to obtain time-dependent groundwater parameters. Additionally, the Gal empirical formula relies on limited parameters and is prone to overextension, resulting in high prediction errors ranging from 84% to 398%. Recognizing that mine water inflow over time reflects the site's hydrogeological conditions, this study introduces the potential line method for water inflow prediction. By calculating the water-rich coefficient based on updated inflow data, the accuracy of predictions improves compared to traditional methods. Using a case study of a single mine, water inflow was predicted using both the water-rich coefficient method and the trend line method. The relative errors for these approaches ranged from 14.2% to 17.3%. Overall, the large well method is best suited for the early stages of mine design and initial production. For ongoing operations, the water-rich coefficient method or trend line method is recommended, provided that water inflow data is collected continuously and comprehensively.