The combinatorial optimization of cross-sections of latticed shells can be solved in a small computational time by using sequential quadratic programming. However, in the case of steel structures, the obtained optimal solution cannot be used as it is because they are usually composed of standardized member cross-sections. In order to solve such a problem, post-processing method to convert the optimal cross-section obtained by sequential quadratic programming to a normalized member cross-section is proposed in this study. It is shown that the proposed method can solve the combinatorial optimization problem for lattice shells in a small computational time.