Abstract: With the rapid expansion of cities and the high-speed flow of residents, the functional configuration of different areas changes over time, and the functionality of urban land shows mixed characteristics. Therefore, exploring the division of urban functional areas and the identification of functional hybridity is of great significance to the rational planning of cities. The traditional functional area division method usually only uses a single data source, which is difficult to fully and accurately reflect the complex relationship within the city. The existing urban functional area division method based on trajectory data and point of interest (POI) data ignores the scale information of different POI data and cannot effectively identify the regional functional hybridity and the time-varying law of regional functional use intensity. A method for urban functional area division and functional hybridity identification based on spatiotemporal semantic fusion model is proposed. Firstly, the urban area is divided into spatiotemporal units based on the spatiotemporal cube model, and the start and end point (OD) matrix is constructed to reflect the spatiotemporal attributes of trajectory data; then the density matrix of POI service population is extracted and standardized to make up for the defect that POI data cannot reflect scale information; then the spatiotemporal semantic fusion model is constructed to effectively fuse OD data and POI data to divide urban functional areas, and reveal the time-varying law of regional functional use intensity with human activities. A spatiotemporal information entropy model is constructed to achieve quantitative evaluation of urban functional hybridity. Finally, an interactive visual analysis system is designed to intuitively understand and interpret the results of urban functional area division and functional hybridity identification through a series of visualization views. Experiments are conducted using real data from Hangzhou, and the effectiveness and practicality of the proposed method are demonstrated through specific case analysis.