Exploring dynamic response-based damage identification under wave excitation is important for establishing a health monitoring system. In our previous study, we analyzed the characteristics of the rigid-body dynamic response and successfully employed them to identify the damage location of a pile foundation. Nevertheless, some recent researches show that the local flexural response resulting from wave impact has stronger energy performance. Therefore, this paper constructed a new damage identification method by making full use of energy and phase slope of the local flexural response. Firstly, a novel frequency domain clustering algorithm based on the Sub-peaks Suppression of Power Spectral Density Method (SSPSDM) was proposed to address the challenge of mode mixing in signal decomposition. Subsequently, the new composite energy factors were developed by combining both energy and phase of the local flexural response to enhance the sensitivity and robustness of damage identification. Finally, the method was validated using the acceleration signals from a laboratorial high-piled wharf model. The results show that the method can effectively extract the damage feature sub-signal and the composite energy factor based on phase slope enables better identification of the damage location and degree, enhancing the accuracy of the damage assessment process.