Abstract: A pilot test on the horizontal well JS6-7P01 in the eastern margin of Ordos Basin achieved a significant breakthrough in the first large-scale hydraulic fracturing of deep coalbed methane (CBM, or coal-measure gas), promoting nationwide exploration and development of deep CBM. This advancement has positioned deep CBM as a critical domain for unconventional natural gas reserves and production growth. However, the mechanistic role of hydraulic fracturing in methane production, beyond enhancing permeability, remains unclear. High costs and operational uncertainties in field trials have hindered the optimization of deep CBM technologies and efficient development. To address this, comprehensive physical simulations, numerical modeling, and theoretical analyses were conducted, including wettability characterization and interfacial molecular mechanism analysis of deep coal, in-situ gas-water occurrence monitoring via online nuclear magnetic resonance (NMR) scanning, realistic macromolecular modeling of deep coal, molecular dynamics simulations of aqueous-phase invasion and displacement, isothermal adsorption experiments under varying moisture conditions, imbibition experiments under different salinities, and scanning electron microscope. These efforts aimed to elucidate the “pressure displacement-imbibition-replacement” mechanism underlying hydraulic fracturing in deep CBM reservoirs. Key findings include:Hydrophilic characteristics: Mineralogical analysis, high-temperature/high-pressure (HTHP) wettability experiments, and macromolecular modeling revealed that deep coal exhibits macroscale hydrophilicity and microscale strong hydrophilicity due to hydrophilic clay mineral filling and oxygen-containing functional groups. Methane adsorption reduction: Isothermal adsorption experiments and HTHP NMR scans demonstrated that fracturing fluids significantly weaken methane-coal interactions. A 1% increase in moisture content reduces maximum methane adsorption by 1.82 m³/t on average. At a 5 MPa gas-liquid pressure differential, fracturing fluids displaced 70.84% of free gas and competitively desorbed 10.42% of adsorbed gas, suggesting higher displacement efficiency under elevated pressure gradients. Nanopore displacement dynamics: Molecular dynamics simulations revealed piston-like displacement in nanopores under high-pressure gradients, converting 97.8% of adsorbed methane into free gas. Capillary-driven imbibition: Sub-5 nm coal matrix pores exhibited steeply increasing capillary forces. For 2 nm pores under reservoir and post-fracturing pressure conditions, strong hydrophilic (30° contact angle) and weakly hydrophilic (70°) pores generated capillary forces exceeding 40.1 MPa and 15.8 MPa, respectively, creating intense imbibition driving forces. Osmotic enhancement: High salinity contrast between fracturing fluids and formation water induced osmotic pressure via chemical potential gradients, enhancing imbibition capacity by ≥46.4%.Desorption kinetics: Bigger coal fragments (1.3 cm vs. 5.0 cm) extended desorption equilibrium by more than 14 times, and decreased recovery by 15.6%.The experimental and simulation results above revealed the mechanism of “pressure displacement-imbibition-replacement” during deep CBM large-scale fracturing, which is a synergistic process combining fracture propagation, permeability enhancement, contact area increasement, microporous connectivity, spontaneous/pressure driven imbibition-replacement, and high-pressure displacement. Four gas production sources and the stag’s contribution were identified, leading to four optimized strategies: dense fracture networks, differentiated well shut-ins to enhance imbibition, wettability modifiers to mitigate water blockage, and controlled-pressure extended flowback. Three future research priorities were outlined. These findings provide critical scientific insights for advancing hydraulic fracturing technologies, optimizing production protocols, and improving well productivity in deep CBM development.