In this research, the principal purpose is investigating the performance of structural elements significantly degrades at elevated temperatures. Steel–concrete composite floor systems are one of the most relevant components in building construction in which fire-induced problems directly damage their performance. The purpose of this study is on employing analytical intelligence technique to predict two major structural characteristics of the steel–concrete composite floor system. Two intelligent methods, ELM-PSO and ELM-GWO, were used as a multi-combination AI method to predict the shear and tensile response of these composite floor systems at high temperature. Accordingly, authenticated data on monotonic loading response of this steel–concrete composite floor system in different heat stages had been employed from different literatures. The results show that ELM-GWO technique presented the best prediction of split-tensile load, and the ELM-GWO provided the best estimation of slip value. By giving each prediction equations, the best equations were proposed. As a result, employing ELM-GWO, successfully presented reliable and decisive results and proved the proficiency of the techniques.