Response surface methodology (RSM) is a statistical method to forecast response values by designing logical experiment programs

adopting quadratic polynomial equation to associate and optimize the relationships between factors and response values. In this work

RSM was applied to optimizing the most common fire resistant additives (APP-MEL-PER) of intumescent fire resistant coating for steel structure. The influence of fire resistant additives and their mutual effects on the fireproof properties of fire resistant coating were discussed. Moreover

relevant mathematical forecast model was also built. The analysis of variance showed that 98.33% response value could be explained by the quadratic polynomial created. The error rate of modeling prediction was only 0.48%. These results proved that this modeling was efficient and precise. The equilibrium temperatures of the optimized sample decreased by 12℃. Besides

the optimized sample had a more uniform foam structure and well-distributed cell sizes

which resulted in better fireproof performance. The optimization for fire resistant coating by response surface methodology was feasible and successful.