Abstract: In order to reduce the error of the empirical formula method in the prediction of the height of the water-conducting fracture zone, and to explore the prediction method of the water-conducting fracture zone with higher prediction accuracy, based on the measured data accumulated in Tengzhou mining area in Shandong, this paper comprehensively analyzes the six factors of mining depth, coal seam thickness, working face inclination length, lithological parameters, roof thickness and inclination angle.With the help of MATLAB data analysis software, the method of principal component analysis is used.Combined with the regression model test and analysis, a regression model for predicting the height of the roof water-conducting fracture zone suitable for the mining area is established, and the results are compared and analyzed with the results obtained by the empirical formula method.The results show that principal component regression model, based on measured data from 12 coal mines in Tengzhou mining area, predicts that the heights of the water-transmitting fracture zones in the two verification working faces of Jiangzhuang coal mine ware 56.239 1 m and 49.102 1 m, respectively, and the relative prediction errors are 3.00% and 2.83% respectively; and the empirical formula method predicted the height of the water-conducting fracture zone of the two working faces are 60.00 m, and the relative prediction errors are -9.89% and -25.65%, respectively.It can be seen that the principal component regression analysis model method is used to predict the water conductivity of the coal roof.The development height of the fracture zone has high accuracy.This model can better predict the development height of the water-conducting fracture zone and serve the prevention and control of water damage in the coal roof.