Abstract: Saturation is an important parameter for gas hydrate resource assessment. The methods for estimating gas hydrate saturation vary greatly according to the differences in morphological characteristics of gas hydrate and its contact relation with sediments. Fracture-filled gas hydrate occurred in forms of a vein, patch, or nodule has been found in the east part of the Pearl River Mouth Basin. However, the gas hydrate occurred as pore-fillings has been identified in the pressure core sample collected from the Shenhu area. The well log responses to fracture-filling and pore-filling gas hydrate are quite different indeed, and they show anisotropic and isotropic appearances respectively. Gas hydrate saturation estimated with anisotropic resistivity model fits well with gas hydrate saturation estimated by pore-water freshening. The gas hydrate saturation near seafloor is about 10% estimated according to anisotropic resistivity. However, it may reach about 60% if estimated from isotropic resistivity which is obviously greater than that estimated from pressure core degassing. Gas hydrate saturations in carbonate layers are similar no matter they are estimated with anisotropic and isotropic resistivity or velocity, which are higher than those estimated from chloride. However, the gas hydrate saturation at this layer estimated from cementation model C1 is near to the value estimated from chloride. The acoustic impedance or P-wave velocity was obtained from the constrained sparse spike inversion. The spatial distribution of gas hydrate saturation can be calculated with linear or iterative forward inversion of rock physics model based on seismic data and well log data. Gas hydrate saturation extracted from the base of gas hydrate stability zone with a time window of 50 ms showed a patchy or nonuniform distribution pattern in the Shenhu area. However, gas hydrate saturation showed linear distribution at the crest of canyons.