This study takes the 2017 “8–8” Jiuzhaigou earthquake as the background, based on the disaster system theory, constructs the Jiuzhaigou scenic disaster resilience evaluation index system from the four dimensions of early warning, rescue, disaster preparedness and management, and analyzes the dynamic evolution characteristics of the scenic disaster resilience in the period of 2015–2024 through the entropy weight-TOPSIS method. The dynamic evolution characteristics of disaster toughness in the scenic area. The study found that: (1) the overall toughness presents the trend of “pre-disaster fluctuation—sudden drop in disaster time—continuous rebound in post-disaster period”, and the toughness index grows at a slower rate before the earthquake, and there is a sudden drop due to the earthquake in 2017; in the post-disaster recovery period, through the ecological restoration, optimization of the emergency channel and the construction of digital twin platform, the resilience of the scenic area is improved. After the disaster, through ecological restoration, emergency access optimization and digital twin platform construction, the toughness has been significantly improved; (2) at the subsystem level, resilience is the fastest-growing due to vegetation gradient restoration and seismic facilities renovation, early warning benefits from AI monitoring network upgrades, and rescue is limited by mountainous terrain; (3) the research deficiency lies in the dilemma of data isomorphism, lack of theoretical applicability, and the standardization of the evaluation system. It is suggested to strengthen the “ecological-facility” composite resilience network, multi-hazard early warning system and community collaborative governance to provide scientific basis for disaster resilience planning in mountainous scenic areas.