Abstract: In view of the problem of difficulty in identifying soft rock interlayers and low efficiency of the traditional drilling method in coal mine roadway roof,an intelligent recognition method based on hole drilling specific power is proposed.By analyzing the mechanical properties such as density,strength,elastic modulus,softening coefficient and others of roof soft rock,sandy mudstone,and fine sandstone in the 3303 working face of Licun Coal Mine,a hole drilling specific power calculation model considering the friction loss of drilling tools is constructed.The BP neural network recognition algorithm with the introduction of momentum factor and adaptive learning rate is adopted.A ZKJM-SRL soft rock interlayer intelligent recognition system is developed,which includes four modules:data acquisition,pre-processing,recognition calculation,and early warning output,achieves real-time acquisition of drilling parameters with a sampling frequency of 10 Hz,data intelligent processing with a delay of less than 50 ms,and automatic recognition of soft rock interlayers.The results show that the recognition accuracy rate of the system for mudstone,sandy mudstone,and fine sandstone reaches 94.2%,91.8%,and 88.5%,respectively,in particular,three soft rock interlayers with a thickness of less than 0.5 m are accurately identified near the FX2 fault zone.