The impervious surface is an important reference index for evaluating the urbanization process and the urban environment. Studying the spatial coverage of urban impervious surface is of great significance for urban ecological environment protection and livable city construction. In this paper, the hyperspectral image of "Zhuhai No.1" is used as the data source, and the land area of Zhuhai City is used as the research area. Support vector machine and random forest machine learning algorithm are selected to combine spectral features to achieve the extraction of urban impervious surface information. The results show that using the hyperspectral data of "Zhuhai 1" to extract the impermeable surface is feasible, and the support vector machine algorithm is more suitable for the extraction of "Zhuhai 1" impermeable surface, with the overall accuracy and Kappa coefficient reaching 92.4% and 0.78; The extraction of urban impervious surface based on "Zhuhai No.1" hyperspectral data can provide reliable theoretical basis and data support for urban planning and construction, analysis of urban land expansion, and assessment of urban disaster risk.