Abstract: Gas control is a critical aspect of coal mine safety, fundamentally relying on the precise understanding of coal fracture evolution and gas seepage characteristics, combined with microseismic monitoring technology to achieve early warning of dynamic disasters. Based on this principle, this study employs raw coal specimens and utilizes a triaxial servo-controlled seepage apparatus to conduct compressive strength tests under different gas pressure gradients. The research systematically investigates the influence of gas pressure on the mechanical response of coal, determines the gas pressure threshold, and reveals the evolution characteristics of gas seepage. The results indicate that, during the initial compaction stage, the primary fractures in the coal gradually close, restricting seepage channels. In the late stage of elastic deformation, the internal pores of the coal sample are further compacted, causing the gas seepage flow rate to drop to its minimum. Near the yield point, the seepage flow rate reaches its lowest value, while the frequency of microseismic events significantly increases. When the coal sample reaches the peak load, the gas seepage channels become instantaneously connected, leading to a sudden surge in flow rate, a sharp increase in microseismic frequency band energy, and the highest number of recorded events. By calculating the entropy values of different wavelet basis functions for microseismic signals, the optimal function (Db6) was selected, and a four-layer wavelet packet decomposition was performed to obtain the normalized energy distribution across 16 frequency bands. The findings indicate that the characteristic frequency bands associated with coal sample failure are primarily concentrated in the 6th to 8th bands, corresponding to a frequency range of 625 Hz to 1000 Hz. This study provides theoretical support and technical guidance for gas control in coal mines and early warning of dynamic disasters.