Abstract: In order to obtain the prediction model of the height of the water-conducting fracture zone applicable to the mining conditions of thick and extra-thick coal seams in Huanglong coalfield, the Jurassic coalfield in Huanglong is taken as the study area, based on 38 groups of measured data of the height of the water-conducting fracture zone, and the impact of mining thickness, slope length of the working face, buried depth of the coal seam, overburden type, and mining method on the height of the water-conducting fracture zone is comprehensively considered, a multivariate nonlinear regression model and a BP neural network model optimized by genetic algorithm for the prediction of the height of the water-conducting fracture zone in Huanglong coalfield are constructed based on the data-driven method, and the two models are applied to the prediction of the height of the water-conducting fracture zone in the 401101 working face of Mengcun Mine. The results show that: The height of the hydraulic fracture zone is affected by many factors. The fitting coefficient of the regression model is increased from 0.52 of the mining thickness to 0.82 of the comprehensive multiple factors. From the perspective of data-driven, it is concluded that the weight value of the factors affecting the height of the water-conducting fracture zone in Huanglong Coalfield is ranked as follows: mining thickness > slope length of the working face > overburden type > buried depth of the coal seam > mining method. The mining thickness and the slope length of the working face are the main control factors affecting the height of the water-conducting fracture zone, which should be controlled during the mining process to achieve the purpose of preventing and controlling the water hazard of the coal seam roof.The application results of the model show that the maximum relative error of the multivariate nonlinear regression model is −3.67%, while the maximum relative error of the GA-BP neural network model is only −1.95%. The relative errors of both prediction models are less than 5%, which can meet the requirements of engineering practice. When the requirements for prediction accuracy are high, the GA-BP neural network prediction model can be selected. The research results can provide a certain basis and reference for the study of the height of the water-conducting fracture zone under the conditions of thick and extra-thick coal seams in Huanglong Coal Field.