In regulated rivers, dams and their reservoirs are the main means to change the flow conditions, and strategic reservoir operations can be an important measure of mitigating ice-jam floods (IJF). Previous studies have shown the effect of single reservoir operations on IJF probabilities, but little is known about the effect of cascade reservoirs on IJF occurrences and severity. This study presents a probabilistic modeling framework that couples river ice and machine learning models to assess the effects of regulated and naturalized flows on IJF backwater levels. The framework was used to assess ice-jam flood hazard along the Sanhuhekou bend, a reach of the upper Yellow River whose hydraulic regime is impacted by eight upstream reservoirs. Naturalized flows and IJF backwater levels will be generated for comparison. Surprisingly, results show that regulation can increase the risk of IJF.