The Chang'E-7 lunar mission, scheduled for launch in 2026, has the primary scientific objective of detecting water-ice deposits within the permanently shadowed regions (PSRs) at the lunar south pole. In this study, we developed a high-fidelity model of the Chang'E-7 Lunar Neutron and Gamma-ray Spectrometer (LNGS) payload using the Geant4 toolkit and established a quantitative inversion relationship between lunar surface water content and epithermal neutron count rates. Through simulations of secondary neutron spectra generated by galactic cosmic ray (GCR) bombardment of the lunar surface, combined with validation experiments using neutron beam calibration at the China Spallation Neutron Source (CSNS), our results demonstrate that: 1) The detector model shows excellent agreement with experimental data, with a relative error of less than 6%; 2) LNGS exhibits significant capability in discriminating soils with varying water content; 3) Within the water-ice content range of 0.01-20 wt.%, the epithermal neutron count rate decreases significantly with increasing hydrogen abundance, following a modified Lawrence model (R²=0.9993). This study provides a robust theoretical framework for interpreting Chang'E-7 orbital data and establishes fundamental technical support for the development of in-situ resource utilization technologies on the Moon.