Bridge structures are required to possess high reliability and robustness against the concurrent effect of extreme loads and environmental attacks. To achieve such interrelated goals, it is necessary to assess the system performance and resilience subjected to multi-hazard impacts and the beneficial effects of any retrofitting or hazard-countermeasure in a lifecycle context. The damaged bridge needs to be restored rapidly over its service life due to the significant economic losses and disruption to transportation networks. For river-crossing bridges, one of the essential hazard mitigation strategies is scour countermeasures. However, a quantitative understanding of the effects of SCs on bridge system resilience is not found. This dissertation presents a critical synthesis of the existing literature that provides relevant knowledge and a profound understanding of probabilistic multi-hazard assessment for bridge structures. Then, a finite element-based probabilistic framework is designed to assess the lifecycle resilience of reinforced concrete river-crossing bridges under seismic, flood-induced scour, and chloride-induced corrosion impacts, including the consideration of a typical scour countermeasure at variable service times. Based on the general performance-based approach, two probabilistic models are formulated, termed the mean-scour fragility analysis (MS-FA) model and the total-scour demand hazard analysis (TS-DHA) model, which produce straightforward functional curves and can be readily used to evaluate the seismic-scour multi-hazard effects. An integrated damage index is defined based on both local and system-level ductility demands to develop a demand hazard model and to estimate the damage-based residual functionality and recovery duration to quantify the lifecycle bridge resilience. Notably, the exceeding probability approach is designed to reveal how progressive and sudden hazards interact and result in resilience degradation and how scour countermeasures contribute to resilience enhancement. The outcomes of the numerical experiment reveal the positive and distinct effects of implementing SCs at different lifecycle intervals. More importantly, resilience time-series demonstrate arbitrary multi-modes and nonparametric patterns. Accordingly, a robust statistical distance-based approach is presented to determine the sequential evolution of time-varying multi-hazard resilience. The proposed framework would assist stakeholders and decision-makers in resilience patterns recognition, assessing the effectiveness of hazard mitigation strategies, and taking short- and long-term proactive intervention actions by specifying resilience thresholds.