<p>Integrating digital and informational technology into geological hazard analysis and early warning represents a transformative shift from a traditional focus on post-disaster relief to a proactive emphasis on pre-disaster prevention. A physical model-driven framework was introduced for the digital simulation of geological hazards, which can be designed to be utilized and interacted with multi-source remote sensing and mapping data in an efficient manner. The framework encompasses disaster scenario element extraction, lightweight processing, and semantic modeling based on various datasets. It incorporates physical–mechanical algorithms to reconstruct the real-time state of hazards, simulate their future evolution paths, and predict post-instability kinematic behaviors. This approach captures the entire evolution of geological disasters. To validate the framework, a case study of a high-elevation landslide in the Xinhua Village segment of National Highway G351 in Ya’an, Sichuan, China was conducted. The dynamic evolution process and staged characteristics of the river-blocking mechanism of the landslide were analyzed from the perspective of energy evolution, reveals that the energy interaction and transfer efficiency between the landslide and the river are critical. Additionally, the ParaView platform was adopted to integrate and visualize model calculation results with digital real-world models. This multi-perspective, multi-dimensional scenario analysis can facilitate a deeper understanding of landslide instability and failure scenarios, demonstrating the practical potential of physics-driven scenario simulation technology.</p>