Abstract: A deformation reconstruction method that introduces transfer learning is proposed and verified by experiment with an unmanned air vehicle (UAV) wing in this paper. A large amount of simulation source data are obtained by modal method first and a small amount of target data are obtained by test of the actual wing. Then the target data are transferred to the source data by the transfer learning algorithm, so that the reconstructed displacement value is closer to the measured value. The results show that this method has greatly improved the deformation reconstruction accuracy of the wing in the actual work environment.