Modeling of high-dimensional aerodynamic data presents a major challenge in the context of aero-loads prediction, aerodynamic shape optimization, flight control and simulation, etc. In this article, a machine learning approach based on convolutional neural network (CNN) is developed to address this problem. CNN is able to implicitly distill features underlying the data, and the number of parameters to be trained can be significantly reduced due to its local connectivity and parameter sharing properties, which is favorable for solving high-dimensional problems in which the training cost can be prohibitive. A hypersonic wing similar to the wing of Sanger aerospace plane carrier is employed as the test case to demonstrate the CNN-based modeling method. First, the wing is parameterized by free-form deformation method and 109 variables incorporating flight status and aerodynamic shape variables are defined as model input. Second, 7431 sample points generated by Latin hypercube sampling method are evaluated by performing computational fluid dynamics simulations based on a Reynolds-Average Navier-Stokes flow solver to finally obtain an aerodynamic database, and a CNN model is built based on the observed data. Finally, the well-trained CNN model considering both flight status and shape variables is applied to aerodynamic shape optimization to demonstrate its capability to achieve fast optimization at multiple flight statuses.