This paper presents a methodology to develop a pyrolysis reaction model that captures the flammability behavior of epoxy-based intumescent flame retardant coatings (EP/IFR) as a function of material composition. This approach adopted Thermogravimetric Analysis (TGA), Differential Scanning Calorimetry (DSC), and Microscale Combustion Calorimetry (MCC) experiments as well as the inverse modeling of the experimental results using ThermaKin2Ds and COMSOL Multiphysics modeling framework. The interactive reactions between EP and IFR were identified and quantified in the resulting lumped model, which is based on fitting of experimental data. It was found that the developed model was able to reproduce the TGA/DSC/MCC data within the prescribed criteria for EP/IFR blends with different material compositions. The model’s extrapolating capability was further validated through accurately predicting the experimental TGA data under different thermal conditions and for blends with different IFR ratios which were not used for model development. Additional TG-FTIR and XPS experiments were conducted to quantify the effect of IFR additive on the evolution of gaseous products and char formation. With a higher ratio of IFR, the amount of fuel dilution gases (i.e., CO2, NH3) was increased and the amount of combustible organic gaseous products was reduced. This work provides the core subset parameters for comprehensive pyrolysis modeling and enables the intelligent design of EP/IFR coatings with enhanced flame resistance.