Extreme heat caused by global climate change has become a serious threat to the sustainable development of urban areas. Scientific assessment of the impacts of extreme heat on urban areas and in-depth knowledge of the cross-scale mechanisms of heat vulnerability forming in urban systems are expected to support policymakers and stakeholders in developing effective policies to mitigate the economic, social, and health risks. Based on the perspective of the human-environment system, this study constructed a conceptual framework and index system of “exposure-susceptibility-adaptive capacity” for urban heat vulnerability (UHV) and proposed its assessment methods. Taking Xiamen City, a coastal metropolis, as an example, spatial analysis and Geodetector were used to explore the spatial and temporal changes, spatial characteristics, and patterns of UHV under multiple external disturbances from natural to anthropological factors, and to reveal the main factors influencing UHV forming and spatial differentiation. Results showed that the exposure, susceptibility, adaptive capacity, and UHV in Xiamen City had a spatial structure of “coastal-offshore-inland”. On the hot day, both the exposure and UHV showed a temporal pattern of “rising and then falling, peaking at 14:00” and a spatial pattern of “monsoonal-like” movement between coast and inland. Coastal zoning with favorable socioeconomic conditions had less magnitude of changes in UHV, where the stability of the urban system was more likely to be maintained. During the hot months, the high UHV areas were mainly distributed in the inland, while coastal areas showed low UHV levels. Further, coastal UHV was mainly dominated by “heat exposure”, offshore by “comprehensive factors”, and inland in the northern mountainous areas by “lack of adaptive capacity”. Multi-scale urban adaptive capacity was confirmed to alter spatial distribution of exposure and reshape the spatial pattern of UHV. This study promotes the application of multi-scale vulnerability framework to disaster impact assessment, enriches the scientific knowledge of the urban system vulnerability, and provides scientific references for local targeted cooling policy development and extreme heat resilience building programs.