A pyrolysis model capable of predicting materials’ fire behavior as a function of concentration was developed for an intumescent flame retardant system: poly(lactic acid) (PLA) blended with melamine (MEL) and ammonium polyphosphate (APP). The model was developed through inverse analysis of data obtained from bench-scale pyrolysis experiments wherein a 0.07-m-diameter disk-shaped sample was exposed to well-defined radiant heating in an anaerobic environment. Sample back surface temperature, sample shape profile and burning rate were measured simultaneously. A numerical pyrolysis modeling framework, ThermaKin2Ds, and a previously developed semi-global thermal decomposition reaction mechanism were employed in the inverse analysis to determine material properties that define the heat and mass transport inside the pyrolyzing solids. The final pyrolysis model was found to predict materials’ fire behavior for a variety of thermal exposures and material compositions. The model construction process revealed that a reduction in gas transfer coefficients helped to reproduce certain features of the burning rates profiles. Idealized cone calorimetry scenarios were simulated to study the influence of additives on materials’ fire behavior, and the results demonstrated that the blend with 5 wt % MEL and 25 wt% APP was found to be the most effective system with a 69% reduction in the average heat release rate comparing to that of PLA. A similar significant reduction has been reported in the literature, supporting the accuracy of this model. This work demonstrates a methodology that enables intelligent design of intumescent flame retardant materials tailored for specific applications, where low flammability is required.