Abstract: Large Ming-Qing ancient building complexes serve as critical material carriers of Chinese civilization, embodying profound historical, cultural, and social value, and have long been operating under conditions of intense public visitation. The superposition of high-density visitor flows with the complex spatial layout, abundant cultural relics, and inherent structural vulnerability of such complexes generates safety risks with pronounced spatiotemporal dynamic characteristics. Although the increasing intensity of open-operation activities has rendered traditional static vulnerability evaluation increasingly inadequate, existing studies on ancient building safety risk assessment have largely focused on individual buildings or small clusters, lacking systematic methods for large complexes containing more than 800 buildings, and rarely couple visitor dynamic effects with static vulnerability in a unified framework. Consequently, this escalating risk landscape underscores the critical need for a comprehensive assessment approach that simultaneously characterizes static vulnerability and dynamic exposure mechanisms. In this context, this paper presents a comprehensive design and application of a coupling-based comprehensive safety risk assessment method tailored for large Ming-Qing ancient building complexes, with a focus on 868 ancient buildings within the Palace Museum. The study begins by analyzing the current state of ancient building risk assessment and the limitations of existing static-only methods to firmly establish the necessity for a coupling-based assessment framework. The proposed method is systematically designed around a “disaster risk framework with static-dynamic coupling” architecture, logically structured into four cohesive components: a Static Vulnerability Evaluation Component that applies the entropy-weight TOPSIS model to seven indicators spanning structural complexity, number of cultural relics, functional attribute, management and security personnel, enclosure-cordon level, security-system level, and fire-protection level; a Visitor-Flow Factor Component that defines seven typical operating scenarios based on 20-year (2000–2019) daily visitor records and China’s statutory holiday regulations, and characterizes visitor-flow effects through a scenario visitor-flow intensity coefficient, a per-building influence weight derived from visitor-aggregation degree and building accessibility, and a building floor-area correction coefficient; a Comprehensive Risk Calculation Component that couples the two through a multiplicative form and obtains four risk grades through dual K-means clustering—classifying the closed-day scenario and the six open-operation scenarios with two separately calibrated threshold sets; and a Response-Type Identification Component that classifies each building’s grade trajectory across all open scenarios. Application of the proposed method to the 868 ancient buildings of the Palace Museum yields concrete and quantitatively verifiable results: the entropy weights identify structural complexity (0.297) and number of cultural relics (0.215) as the dominant factors of static vulnerability; the 868 buildings are systematically classified into three response types based on the trajectory of their risk grade across the six open-operation scenarios—persistently in Grade I–II (Stable High, 357 buildings, 41.1%), grade-elevating across the Ⅰ–Ⅱ / Ⅲ–Ⅳ boundary (Quickly Escalating, 94 buildings, 10.8%), and persistently in Grade III–IV (Stable Low-Medium, 417 buildings, 48.1%); the quickly-escalating buildings represent the new category of buildings that dynamic assessment is able to identify but static assessment alone cannot. To validate the practical value of the method, three representative regions corresponding to dense large halls with peak visitor flow (SDD), nested courtyards with dense exhibited cultural relics (ZBG), and closed-to-public offices and storage (NSS) are selected as case studies, and differentiated countermeasures are designed and deduced for each, achieving regional average comprehensive risk reductions of 15.62, 12.43, and 3.54 percentage points respectively, with the regional average risk grade dropping from Grade I to Grade III for SDD, from Grade II to Grade III for ZBG, and remaining at Grade III for NSS. In conclusion, this paper presents a robust and detailed methodology for the comprehensive safety risk assessment of large Ming-Qing ancient building complexes, providing methodological support for the shift from single-grade control toward typological control that simultaneously considers static vulnerability and dynamic exposure mechanisms, and serving as a quantitative basis for advancing the differentiated safety management of cultural heritage sites worldwide. Future work will be directed toward incorporating real-time visitor flow monitoring for adaptive risk evaluation and extending the framework to multi-hazard simulations beyond visitor-driven safety risks.