微震监测技术在国内外煤矿和非煤矿山中已经广泛使用，对预测岩爆、冲击地压、防治水害、深部采场稳定性及顶板冒落等突发灾害起到了重要作用。本文基于邯邢矿区九龙矿15249N工作面微震监测结果，提取其中10种微震属性数据，通过对微震属性的优化和误差分析，优选出矩震级、滑动位移、体变势、能量、静态应力降等5种震源参数，结合巷道揭露和钻孔约束方法进行BP人工神经网络方法计算，通过穷举式搜索（ES）算法和试凑递增法得最优属性顺序和隐含节点数，建立一种基于BP神经网络的微震多属性煤层厚度预测模型。对该预测模型进行误差分析和相似区域判定，结合实际地质条件验证得出该模型在工作面煤厚判定中具有良好的应用效果。

Microseismic monitoring technology was widely used in coal mines and non-coal mines in china and abroad, which played an important role in predicting rock burst, rock burst, prevention of water damage, deep stope stability and roof caving and other sudden disasters. Based on the microseismic monitoring results of No.15249N Face in Jiulong Mine of Hanxing mining area, this paper extracted ten kinds of microseismic attribute data. Through the optimization and error analysis of microseismic attributes, five kinds of source parameters including moment magnitude, sliding displacement, volume change potential, energy and static stress drop, were optimized. Combined with roadway exposure and drilling constraint method, BP artificial neural network method was used to calculate. The optimal attribute order and hidden node number were obtained by exhaustive search ( ES ) algorithm and cut and trial increasing method. A microseismic multi-attribute coal seam thickness prediction model based on BP neural network was established. The error analysis and similar area determination of the prediction model were carried out. Combined with the actual geological conditions, it was verified that the model had a good application effect in the determination of coal thickness in the working face.