The use of metaheuristic optimization algorithms in building structure design is increasingly emphasized and developed. These algorithms have significant advantages over traditional gradient-based methods due to their ability to find globally optimal solutions, particularly for complex, highly nonlinear design challenges. In this study, metaheuristics are employed to automatically find the optimal design for steel frame structures in industrial buildings. This optimization problem is rarely addressed in common practice in Vietnam. The objective is to identify the optimal cross-sectional dimensions of the structural members to minimize the total weight while ensuring safety and serviceability according to design standards. All combinations of impacts on the structure, including dead load, live load, wind load, and foundation settlement, are fully considered. Popular metaheuristic algorithms such as Differential Evolution, Jaya, and Rao are utilized. The resulting numerical calculations are compared to assess the performance of each algorithm. The study demonstrates that applying metaheuristics to optimal design offers advantages in terms of time and quality of design solutions compared to the traditional trial-and-error method still used by engineers.