Industrial steel buildings are widely used in construction projects due to numerous advantages such as cost savings, fast construction time, large span capability, and better suitability for large-scale production and storage purposes. Due to the large span characteristics, the column and beam sections in industrial steel buildings often use composite steel, where the dimensions of the sections are reasonably designed to ensure both structural performance and cost efficiency. However, with such complex design variables, selecting the appropriate sizes poses a challenging design problem. In this paper, we introduce a modern method to address this issue by setting up an optimization problem for determining the optimal cross-sectional sizes of industrial steel frame buildings. The goal is to minimize total weight while ensuring safety and usability according to design standards. The constraints are formulated according to TCVN 5575:2012 standards. The design variables are treated as discrete variables, aligning with practical production and construction realities. An improved differential evolution algorithm is employed to solve the optimization problem. Calculation results demonstrate that the optimization method not only saves time but also enhances the quality of solutions compared to the trial-and-error method commonly used by engineers.