Abstract: Based on enhanced variational mode decomposition and correlation analysis, this paper develops a novel denoising method for improving analytical accuracy of swing signals generated from hydro-electric generating units with strong noisy background and complicated electromagnetic interference. First, we construct the Hankel matrix of a swing signal and calculate its components via singular value decomposition. Then, we single out an effective singular value using the mean filtering strategy, obtain the Hankel estimation matrix, and inversely reconstruct the signal, thereby achieving an enhancement of those signals in the low band. The signal so reconstructed can be decomposed into a collection of mode components by variational mode decomposition, and thus using autocorrelation analysis, we can calculate the energy focusability indexes of the normalized autocorrelation functions for all its mode components. Finally, by these indexes the effective components are selected and aggregated into a denoised swing signal. Application to denoising the simulated signals and the swing signals measured from a hydro-electric generating unit shows that our method is capable of achieving a good denoising performance.