To deal with the issues of slow network training speed, large number of fuzzy rules and insufficient accuracy of traditional error calibration methods, a calibration method based on Principal Component Analysis and Self-Construction Fuzzy Network (PCA-SCFN) is proposed in this paper to realize real-time high-precision deformation reconstruction of integrated wing antennas. Firstly, a displacement-node degree of freedom error model is established based on the inverse finite element method (iFEM), and the inverse problem is solved by the monotone fast iterative shrinkage thresholding algorithm (MFISTA). Secondly, the PCA dimensionality reduction method is introduced to simplify the training network complexity by reducing strain dimensions. Thirdly, non-uniform rational B-spline (NURBS) fitting is applied to the small-sample training set to expand the data, enhancing network generalization and reducing the influence of noise on the training set. Finally, the SCFN is trained based on triangular membership functions (MF) and Takagi-Sugeno (T-S) fuzzy model to obtain the fuzzy rules. The results of wing loading experiment show that the PCA-SCFN-based calibration method can greatly improve the reconstruction accuracy, and at the same time, it has faster training speed and fewer rules. For a load of 80 N, the maximum reconstruction error is only 0.46 mm when the maximum deformation of the structure is -134.36 mm, and the training time of the SCFN is only 9.714 s, and the number of rules is only 121 at most. Therefore, the calibration method based on PCA-SCFN is an effective approach that can be applied to wing deformation monitoring.