Abstract: An optimal control model of seats based on estimation of distribution algorithm (EDA) is proposed. Firstly, the probability model of individual distribution in solution space is established by statistical learning. New populations are gotten by sampling the probability distribution randomly. The algorithm is iterated to realize the evolution and finally to get the best individuals. The algorithm is compared with genetic algorithm (GA) through simulation experiments. The experimental results show that the estimation of distribution algorithms can quickly obtain a satisfactory solution in solving multi-leg seat allocation problem, and the solving speed is 6 times as fast as that of genetic algorithm.