Under extreme disasters, networked hybrid AC/DC microgrids may suffer internal dependent electrical failures caused by network topology changes and power flow redistribution following external independent mechanical failures. As a result, component fault evolution and system dynamic response exhibit significant cross-timescale coupling, which poses challenges for existing resilience assessment methods in consistently characterizing the full fault evolution process and identifying critical risk moments. This paper proposes a resilience assessment method based on the ORNL-PSerc-Alaska (OPA) model, considering the multi-timescale interwoven effects of component fault evolution and system dynamic response. First, a fault modeling framework is developed to represent external independent mechanical failures and internal dependent electrical failures with spatiotemporal correlations, thereby characterizing the chain evolution of component failures under extreme disasters. Then, a multi-timescale assessment method is established to capture the interwoven relationship between component fault evolution and system dynamic response. Furthermore, a critical state index oriented toward power supply assurance for important users is proposed to identify key risk moments during fault propagation. Finally, the effectiveness of the proposed method is validated on a modified IEEE 33 bus test system for networked hybrid AC/DC microgrids.