Spontaneous heating in the active goaf area during normal mining processes poses increased threats to mine productivity and safety, as evidenced in events induced by spontaneous combustion of coal. To control and mitigate this engineering problem, there is a need to gain critical knowledge of spontaneous combustion in the longwall goaf area, which can be achieved through a combination of field tests and numerical modeling. This paper introduces the spontaneous combustion management system widely used in Australia and presents Computational Fluid Dynamics (CFD) models for the simulation of gas flow dynamics in the goaf area, based on the site conditions of an underground coal mine where coal seam gas is predominantly comprised of carbon dioxide. The models were validated with gas monitoring data and used to conduct parametric studies for proactive goaf inertisation optimization. Qualitative and quantitative analysis of simulation results indicated that better goaf inertisation could be achieved when nitrogen was injected via cut-through at 250 m on the maingate (MG) side and surface boreholes at 100 m and 700 m on the tailgate (TG) side, with a total injection rate greater than 1750 l/s. The oxygen concentration on the MG and TG side dropped below 5% at distances of 120 m and 75 m behind the longwall face, with an oxidation zone area of 35375 m2, which was approximately one-third of the oxidation zone area of the scenario without inert gas injection. Simulation results help shed light on improving current goaf inertisation practices to effectively reduce the risk of heating in goaf areas and improve mining process safety based on Australian conditions and practices.