<p>In recent years, marine heatwaves have become stronger and more frequent, driving more severe coral bleaching. To identify impacted coral reefs, bleaching prediction models based on Degree Heating Weeks (DHW) are widely used. Their accuracy, however, is mixed as they rarely account for impacts of other ecological drivers. We developed a multi-parameter Bayesian spatiotemporal model for Australian coral reefs considering ecologically relevant variables that can act in conjunction with thermal stress to modify bleaching. DHW remained the dominant driver of bleaching risk, but its effect was strongly modified by environmental context. The clearest interaction was with ONI (Oceanic Niño Index), indicating stronger bleaching sensitivity under El Niño-like conditions. On the other hand, bleaching sensitivity to DHW weakened with increasing distance from land. Depth, turbidity, and long-term thermal variability were negatively associated with bleaching probability. Under a standardized high-thermal-stress scenario, the model predicted strong spatial heterogeneity in bleaching risk, with some reefs retaining relatively low probabilities of severe bleaching and potentially representing candidate refugia. A qualitative comparison with the reported 2024 Great Barrier Reef bleaching pattern further suggested that the multi-parameter model better distinguished low or moderate bleaching from more severe bleaching states than the DHW-only model, especially in the far northern Great Barrier Reef, where it was less likely to overpredict bleaching risk. These results show that incorporating environmental context and spatial dependence improves bleaching prediction beyond temperature-only models and provides a more spatially differentiated basis for reef monitoring and management.</p>