Abstract: The core task of safety evaluation for coal seam floor mining under water pressure is to assess the risk of floor water inrush. Traditional water inrush risk evaluation often calculates the water inrush coefficient based on static Ordovician limestone water levels, so the resulting safety evaluation of coal seam floor mining under water pressure lags behind changes in water inrush risk and cannot reflect the future risk of floor water inrush during mining. To address this problem, a safety evaluation method for coal seam floor mining under water pressure based on dynamic prediction of Ordovician limestone water level was proposed using mining area Ⅱ0316 of Inner Mongolia Limin Coal and Coke Co., Ltd. as the engineering background. First, a water level variation curve was drawn according to observed Ordovician limestone water level data from the mine, and a Logistic curve was introduced to fit the observed data and predict the Ordovician limestone water level during future mining. Then, based on the predicted Ordovician limestone water level, the five-map and two-coefficient method was used to draw contour maps of failure depth of the coal seam floor protective layer, thickness of the coal seam floor protective layer, effective protective layer thickness, and water pressure borne by the coal seam floor protective layer, and the water inrush coefficient and mining-under-pressure coefficient were calculated. Finally, safety evaluation of mining under water pressure was carried out according to the three-level discrimination standard, and a comprehensive zoning map of safety for mining under water pressure was drawn to divide safe areas for mining under water pressure and water inrush risk areas. The results showed that the coefficient of determination of the Logistic fitting curve was greater than 0.995, indicating that the curve could accurately reflect the water level evolution of the Ordovician limestone aquifer under continuous recharge conditions. The safety evaluation results of mining under water pressure obtained based on dynamically predicted water levels could prospectively reflect the risk of floor water inrush during future mining, solving the problem of insufficient timeliness in evaluation based on static water levels.