A method is proposed to characterize the fast neutron scattering cross-section (σf) quantitatively by the combination of inelastic gamma rays and captured gamma rays, so as to realize the gas saturation evaluation of CO2-injected heavy oil reservoirs based on the three-detector pulsed neutron logging technology. Factors influencing of the evaluation effect of this method are analyzed and the effectiveness of this method is verified by a simulation example. By using the Monte Carlo simulation method and the physical model of bulk-volume rock, the relationship between σf and CO2 saturation is studied, and the saturation interpretation model is established. The influences of formation temperature and pressure, heavy oil density, borehole fluid and reservoir methane content on the evaluation results of CO2 saturation are analyzed. The results show that the characterization of σf by the combination of secondary gamma information can eliminate the influence of formation lithology, borehole fluid and methane content are the main factors affecting the quantitative monitoring of CO2 saturation, and the effects of formation temperature and pressure and heavy oil density are negligible. The simulation example verified the feasibility of the method for evaluating the CO2 saturation of CO2-injected heavy oil reservoirs.