Microbial decomposition of carbon and biogenic methane in coal is one of the most important issues in CBM exploration. Microbial C-N-S functional genes in different hydraulic zones of high rank coal reservoirs was studied, demonstrating high sensitivity of this ecosystem to hydrodynamic conditions. The results shows that hydrodynamic strength of the 3# coal reservoir in Shizhuangnan block gradually weakened from east to west, forming the transition feature from runoff area to stagnant area. Compared with runoff area, the stagnant area has higher reservoir pressure, gas content and ion concentrations. The relative abundance of genes associated to C, N and S cycling was increased from the runoff area to stagnant area, including cellulose degrading genes, methane metabolism genes, N cycling genes and S cycling genes. This indicates that the stagnant zone had more active microbial C-N-S cycle. The machine learning model shows that these significantly different genes could be used as effective index to distinguish runoff area and stagnant area. Carbon and hydrogen isotopes indicate that methane in the study area was thermally generated. The methanogens compete with anaerobic heterotrophic bacteria to metabolize limited substrates, resulting in low abundance of methanogens. Meanwhile, the existence of methane oxidizing bacteria suggests biogenic methane was consumed by methanotrophic bacteria, which is the main reason why biogenic methane in the study area was not effectively preserved. In addition, weakened hydrodynamic conditions increased genes involved in nutrient cycling contributed to the increase of CO₂ and consumption of sulfate and nitrate from runoff area to stagnant area.