Abstract: The height of the water-conducting fracture zone, as a hidden disaster-causing factor in coal mines, is an important indicator for assessing the degree of roof water hazard. To address the issue of accurately predicting the height of the water-conducting fracture zone in mining overburden rock, 50 sets of measured data on the height of water-conducting fracture zones from coal mines across the country were collected. Key influencing factors such as mining thickness, hard rock ratio coefficient, working face length, and mining depth were selected. A fuzzy clustering analysis method for the height of the water-conducting fracture zone was established. By calculating the membership degree of the clustering analysis, a threshold λ of 0.94 was set as the standard, and the heights of the water-conducting fracture zones in the selected coal mines were classified into 22 categories. Combined with a neural network algorithm, the measured data of the water-conducting fracture zone heights from three coal mines were trained, and predicted values were obtained. The research results showed that the relative errors between the predicted and measured values of the water-conducting fracture zone heights for the three coal mines ranged from 4.1% to 6.6%. This accuracy is significantly higher than that obtained from multi-parameter fitting regression formulas and the recommended calculation methods in the "specifications." This demonstrates that the prediction model based on the combination of fuzzy clustering analysis and BP neural network has high accuracy and applicability in predicting the height of the water-conducting fracture zone, providing a theoretical basis for safe coal mining and the prevention and control of roof water hazards in working faces.