This study aims to build an integrated Bayesian Network (BN) model to assess the health of freshwater ecosystems in South Korea’s Nakdong River Basin by uncovering complex causal relationships among key environmental stressors. The BN framework includes multiple variables, such as drought indices, hydrological metrics, meteorological conditions, and water quality indicators, to evaluate their combined effects on biological health indices. The results show statistically significant dependencies between ecosystem health and hydro-climatic stressors, with drought indices exerting strong influence in specific sub-basins. Scenario-based simulations enable probabilistic predictions of biological responses to changes in environmental conditions, allowing for quantitative assessment of system sensitivity. This approach provides a structured understanding of variable interactions and offers early warning insights into ecosystem degradation. The findings support using BN models as decision-support tools for adaptive ecosystem management, long-term monitoring, and restoration planning under future climate variability. Ultimately, this study contributes to developing science-based strategies for sustainable governance of riverine ecosystems.