Abstract: In order to make up for the insufficiency of researches on the structural wind-resistant optimization of large span roofs, the quantum-behaved particle swarm algorithm was adopted to optimize a large-span roof under wind loads. Equivalent static wind loads on the structure were obtained from the database of wind tunnel tests, and the search space of discrete variables was formed according to the shape steel table. An improved fitness function was proposed by defining a coordinate factor for constraint violation, and the over-flow dealing technique was constructed to guarantee the feasibility and convergence of the optimization. Ten runs of computation were carried out to determine the optimal design of the portal frame, and the validity of the optimal solution was checked in all wind directions. The research results show that the objective function decreases monotonically with respect to iterations, standard deviation for total mass is only 4% of its mean value, and the mean iteration number for every run is 24. Therefore, the quantum-behaved particle swarm algorithm displays strong robustness and high efficiency in the wind-resistant optimization of portal frames.