Abstract: As an important unconventional hydrocarbon resource, volcanic reservoirs hold significant potential for exploration and development. However, their complex lithology, low porosity and permeability, and strong heterogeneity in fluid distribution present substantial challenges for fluid identification and prediction. To enhance the accuracy of fluid prediction in volcanic reservoirs, this study proposes a fluid bulk modulus dispersion attribute inversion method (FS-AVO) based on a frequency scanning mechanism. The method leverages the dispersion and attenuation properties induced by seismic wave propagation to enhance the accuracy of fluid prediction in volcanic reservoirs. First, a volcanic rock physics model with partially saturated fluids is developed to describe poroelastic behaviors of volcanic reservoirs, where the dispersion of seismic waves is attributed to wave-induced fluid flow effects, as described by the patch saturation theory. Rock physics modeling indicates that the fluid bulk modulus is more sensitive to gas saturation than the P-wave velocity within seismic frequency range. Building on this, the study derives an expression for the frequency-dependent reflection coefficient related to the fluid bulk modulus and develops a dispersion attribute inversion method based on the frequency scanning mechanism. Synthetic tests show that the fluid bulk modulus dispersion attribute D_Kf-max responds more sensitively to changes in gas saturation than D_Kf, as calculated by the conventional dispersion property inversion method (FD-AVO). Application to field data further confirms that the calculated D_Kf-max significantly improves the prediction accuracy of volcanic gas reservoirs and enhances the ability to characterize fluid distribution in complex reservoir settings.