Abstract: The determination of coal floor damage depth is an important decision factor of whether the water-bursting happened in coal seam floor, so the damage depth prediction of coal seam floor in coal mine deep mining has an important meaning for dangerous degree discrimination of coal floor confined water burst. Based on the BP neural network, and considering all factors that effect floor damage, through the analysis and processing of a large number of samples data, the paper predicted the floor damage depths of 21201 working face, 21301 working face and 2408 working face in Chensilou Coal Mine were 11.67 m, 13.61 m and 14.1 m. Finally, the paper got forecasting formula between floor damage depth and mining depth.