A new damage identification method was proposed to identify common damages of rotor blades. Firstly, the method performed a variational mode decomposition (VMD) on the vibration signals collected by wireless measurement and control system, obtaining a series of principal modes in different frequency bands. Then the cross-sample entropy (CSE) was calculated between different sensors in the same frequency band. Finally, the entropy values, which were damage features, were taken input to a SVM optimized by the empire competition algorithm （ECA）. By analyzing the experimental data, the results show that the feature based on cross-sample entropy has a high degree of discrimination. Damages with different positions and different sizes were identified using this method, and the total identification accuracy is 98.67%, which proves the effectiveness of the proposed damage identification method.