Global warming is projected to shift high-impact flood risks induced by rain-on-snow (ROS) spatially and temporally. However, decision-makers lack guidance on how the events resembling historic ROS flood extremes with destructive consequences would unfold in a warmer climate. This study examined warming-induced change in four historically significant, decision-relevant ROS flood extreme events over the Continental United States using a storyline approach. Specifically, we performed 1-km hindcast and five counterfactual delta-warming (with an increase of air temperature from 1K to 5K with a 1K increment) simulations of the four events using the state-of-the-science Energy Exascale Earth System Model (E3SM) Land Model (ELM) after model calibration. The basin-wide terrestrial water input (TWI) for different extreme events could show increasing, decreasing, or non-monotonic trends with warming, due to changes in rainfall and snow contribution. Runoff shows divergent responses to warming for different events due to changes in TWI and soil infiltration caused by altered soil moisture and thawed soil. High elevations with deep snowpacks tend to have significantly increasing runoff in a warmer climate. Furthermore, we used the ELM-derived hourly 1-km runoff forcings to drive the 2-dimensional hydrodynamic Overland Flow Model for simulating overland and channel flow. Driven by changing runoff, the warming-induced change of maximum inundation depth shows heterogeneous spatial patterns. As the warming level increases, the maximum inundation depth can increase by more than 1-m in some regions. Our study demonstrates that ROS flood extremes could intensify with warming, highlighting the need for a warming-adaptive, event-customized strategy for flood control and water management.