Abstract: As far as now, a large amount of the address data have been collected by the city governments. However, these data are managed and constructed by different departments in the city. Moreover, the construction of the address data are not from the source of geographical entity. Therefore, these issues result in the failure of intensive construction and unified sharing of these data. In this study, we try to set up a multi-level address semantic model and address repository based on application-oriented refinement management requirement, which links the repository to the geographic entities. The constructed address semantic model and repository generate a unique geographic base from the address to the building entity at the building or unit levels. It can also provide the comprehensive support on wisdom city construction, operation and management. It also has been widely applied in COVID-19 epidemic prevention and control.

Key words: address model, address base, geographical entity association, practical application, epidemic prevention and control