Making inexpensive and efficient is one of the main requirements in the structural design process. In recent decades, the optimization algorithms have been developed for use in the engineering sciences. Optimization algorithms based on the natural phenomena are most useful criteria in structural design. In engineering, truss is commonly used in steel structures. In this paper, Teaching-Learning-Based Optimization (TLBO) algorithm has been used to optimize the steel truss. TLBO can calculate accurate and optimum value of a functions, using only the common controlling parameters such as the mean value in each iteration. Four steel trusses have been analyzed to investigate how TLBO improves designing process of the steel trusses. The results shown that TLBO algorithm has satisfactory performance with less calculation than other optimization algorithms.