Multi-point aerodynamic optimization of a transonic wing using data mining is discussed. Design problem has two objectives which are minimization of drag coefficient at Mach number 0.6 and 0.8 respectively. Here, Mach number 0.6 is considered as a subsonic condition, and Mach number 0.8 is considered as a transonic condition with the local shock. To reduce the local shock that causes wave drag, the sweep back angle is required in transonic condition. On the other hand, the sweep back angle reduces lift to drag ratio in subsonic condition. Thus, a complex high lift device like a flap is required. Moreover, the torsion at wing root becomes stronger with high sweep back angle. As a result, the wing structure weight becomes heavy. To design high efficient new generation civil aircraft, the design knowledge which implements a subsonic and a transonic aerodynamic performance simultaneously with few structure penalty is expected. In this study, tapered wing geometry is defined with two cross sections. 31 sample designs are calculated by the unstructured Euler solver and Kriging surrogate models for the resulting drag coefficient of subsonic and transonic condition are constructed. Using these models, nondominated solutions are obtained by genetic algorithm (GA). Analysis of variance (ANOVA) and Self-organized map (SOM), which are data mining techniques, are also applied to obtain the relationship between design space and solution space. According to this result, there is trade-off between two objective functions and compromised design can be considered. According to data mining result, there is possible to find the design which achieve low drag with low sweep back angle and contrived cross sections.