Probable Maximum Flood (PMF) estimates are central to dam safety assessments but are commonly derived through deterministic workflows that provide limited insight into the influence of hydrological modeling assumptions. While uncertainty in extreme meteorological forcing is widely acknowledged, uncertainty arising from the rainfall–runoff transformation is often embedded implicitly and rarely explored explicitly. This paper presents an ensemble based hydrological framework to investigate the propagation of hydrological uncertainty in PMF estimation while treating meteorological forcing as fixed. The framework combines multiple conceptual rainfall–runoff models, several parameter sets to account for equifinality, and multiple baseline hydrological conditions representing antecedent basin states. The resulting ensemble produces distributions of PMF inflow hydrographs, peak flows, and flood volumes. Results show that hydrological uncertainty alone can generate substantial variability in PMF estimates under identical extreme forcing. Beyond uncertainty quantification, the framework is used diagnostically to examine the sensitivity of PMF outcomes to key hydrological assumptions, supporting conservative but explicit and traceable PMF interpretation. This approach enhances the robustness and transparency of PMF assessments and aligns with evolving dam safety practices within the CDA community.