The common surrogate models used in aerodynamic optimization are mostly established based on empirical risk minimization, which have two problems: firstly, the precision of prediction is highly dependent on sample population size, however the computational cost is large by training samples with computional fluid dynamic (CFD) method, secondly, the over fitting problem can not be avoided due to reducing training error blindly. In order to solve the problems, the idea of building a surrogate model based on the principle of structural risk minimization is proposed, and a new kind of surrogate model is established through support vector regression (SVR). The model uses kernel function mapping the practical problems to high dimensional feature space, and reduced the Vapnik-Chervonenkis (VC) dimension by using linear discriminant function instead of non-linear discriminant function. The method has better generalization ability and can avoid the complexity for high dimension problem, which is proved by comparing with other surrogate model in case of little samples. An aerodynamic optimization is done for transport’s wing based on SVR surrogate model, it indicates that the surrogate model by SVR method can predict the aerodynamic characteristics quickly and accurately, by which the efficiency of the aerodynamic optimization can be improved, and the optimal results are reliable and controllable.