In Unmanned Aerial Vehicle (UAV) developments, one of the demands is to design a UAV that has optimal dynamic stabilities. The implementation of a wing with appropriate geometries has a significant effect on the dynamic stabilities of UAVs. The aim of this study is to predict the dynamic stabilities of UAVs based on wing geometries. The wings geometries considered are dihedral and tip-twist angles, while the dynamic stabilities predicted are short periods and Dutch roll. Both dynamic stabilities were defined by the related eigenvalue which was obtained by XFLR5 numerical software. The relationships between wing geometries and dynamic stabilities were obtained using the Artificial Neural Network (ANN) method. Based on the ANN training, the eigenvalues of short period and Dutch rolls as the representation of the UAV dynamic stabilities can be predicted by using 2-9-3-1 and 2-6-5-1 of ANN structures, respectively, with the mean square error of 1.0712e–07 and 2.6170e–08, respectively.