Canoeing is an important Olympic event; however, high-precision positioning and map data visualization technologies have not been widely adopted during the training phase of canoeing. To bridge this gap, this study proposed a domestically pioneering canoeing sport monitoring system based on real-scene 3D. This system integrates high-precision positioning, virtual reality (VR), and virtual-real fusion technologies, providing athletes and coaches with a straightforward, precise data analysis platform. First, this study presents an overview of the background and significance of the construction of the system. Then, it describes the architecture, major functions, and database design of the system. Finally, it introduces the software system development using technologies including the Cesium platform for 3D geospatial applications, ArcGIS Server, ArcGIS API for JavaScript, and WebSocket. The methods for developing key functions are also described. The core functions encompass 3D visualization of the training field, venue query and positioning, virtual-real integration of trajectories, real-time positioning and monitoring, trajectory playback, and data analysis. Therefore, this system enjoys the advances of high-precision positioning, 3D real scene visualization, and virtual-real fusion. The applications of this system will enhance the efficiency and quality of canoeing training and provide a valuable reference for related research fields.