The global 30m resolution artificial land cover dataset (2010) is a dataset developed by a team led by the National Basic Geographic Information Center from 2009 to 2014. This data set is a part of the global 30m resolution land cover data set (2010), which reflects the land types of human living, industrial and mining, commerce, transportation, etc., and the carriers of human living places, operating spaces and projects. The spatial distribution of artificial land cover types objectively reflects the resource allocation, industrial layout and economic situation of human society. The dataset discussed in this paper relies on the global land cover remote sensing mapping project, including 9907 Landsat TM5, ETM multispectral image data and 2640 environmental disaster reduction satellite (HJ-1) multispectral image data. In the process of land cover classification information extraction based on satellite remote sensing data, the identification and information extraction of artificial land cover types are the most complex. The main reason is that the artificial ground cover is mostly distributed in the form of composite pixels, inlaying and interlacing, fragmented and scattered. The research and development of the global 30m resolution artificial surface data set (2010) is based on the diversity and complexity of spectral and morphological features reflected by the artificial surface coverage on satellite remote sensing images, and the comprehensive use of existing classification technologies to organically combine pixel classifiers, object segmentation and knowledge-based discrimination. The POK (Pixel Object Knowledge) classification method based on "pixel object knowledge" is adopted.