Cascading natural hazards pose growing threats to global communities, as seen in recent earthquake-triggered landslide–flood sequences. Traditional single-hazard models fail to capture complex multi-hazard interactions and temporal dependencies that amplify disaster impacts. This session showcases cutting-edge advances in cascading hazard modelling, including machine learning, field–model integration, and cross-scale techniques. Key themes include multi-physics modelling of hazard chains, real-time prediction systems for cascading events, climate-driven shifts in hazard dynamics, and urban-scale applications of integrated models, with case studies highlighting earthquake–landslide chains, volcanic–lahar events, and compound flooding scenarios. This interdisciplinary session integrates multiple subthemes across the conference, bringing together those seeking to move beyond fragmented models toward holistic cascade modelling for improved disaster preparedness and resilience planning.