River management faces environmental and social changes; therefore, applying information and communication technologies in the construction domain is required. Since most of the data collected and stored during the operation and maintenance of rivers is machine-unreadable, much effort and time are required to find and retrieve the necessary information and to associate the static data of facilities with digital real-time data. Digital twins are expected to increase infrastructure management efficiency and quality; however, applying digital twins to river management is difficult because most of the data are machine-unreadable. Therefore, accumulating machine-readable data and associating static data with dynamic data are required. To meet the requirement, this study proposed a product model of the river domain (referred to as the Industry Foundation Classes (IFC) River) and data flow for static and real-time data. This study broke down the elements of rivers into spatial and physical elements and developed classes for the IFC River. The IFC River includes rivers, river channels, and facilities for river management. The structure of the IFC River follows that of other domains of the IFC so that the IFC River can describe complex fields that include multiple domains. As Rivers consist of linear structures (e.g., river channels, levees) and complex structures (e.g., sluices, hydraulic gates), IFC River describes the shape of linear structures using IFC Road, an IFC schema for the road domain. Besides, the data described by the IFC River are static, whereas the digital twin utilizes dynamic real-time data. This study also proposed a data flow that uses a game engine, 3D modeling software, libraries for handling IFC, and NoSQL to realize the digital twins, which contain shapes, properties, and real-time data. IFC River and the data flow this study developed could describe the rivers, the river management facilities, and real-time data as machine-readable data.