In order to coordinate the land use contradiction between living,production and ecological space, and solve the problems such as inaccurate discrimination and insufficient data coverage in the identification of urban "living production ecology space" by datadriven method, a method that can accurately identify the function of " living production ecology space" was provided. Based on the data-driven method, combining the multi-feature information of multi-source heterogeneous data such as POI(point of interest), AOI(area of interest) and remote sensing, the recognition accuracy and scale effect of different data sources were analyzed when they were used as feature factors in functional evaluation system and classification model. Taking Wuhua District of Kunming as the experimental object, the research results show that the recognition accuracy based on multi-source geographic data reaches 92% and 94%. The multifeature information of multi-source data can significantly improve the identification accuracy of urban functional areas, provide a new method for accurate identification of urban functional areas, and provide data and method support for territorial spatial planning on a smaller scale.