Abstract: Automated identification of structural modal parameters is an important area of structural dynamics. Traditional time-domain identification methods have the problem of inefficiency when dealing with large amounts of response data. This study integrates the Dynamic Mode Decomposition (DMD) technique and the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, and proposes a rapid automatic identification method for structural modal parameters based on structural free decay response. The automatic identification strategy for structural modes is proposed based on the rank stability of physical modal parameters in DMD. By evaluating the silhouette coefficient (Sil) and the number of candidate physical mode clusters, the key hyperparameters of DBSCAN is automatically determined by observing significant changes in the silhouette coefficient, leading to the accurate identification of structural modal parameters. The accuracy of this method is demonstrated by comparing with Finite Element Method (FEM) using a simply supported beam model. Additionally, the capability to quickly identify modal parameters in large datasets is illustrated by comparing to the Eigen Realization Algorithm (ERA) on an inhomogeneous cantilever plate model. Finally, the practical application is validated by using a physical wing model, where structural responses are measured by a non-contact motion capture system. Our modal identification results agree well with those from the LMS. Test Lab modal testing system. The noise resistance of this method is demonstrated on the physical and FEM model of the wing under 10% noise conditions.