Industrial lands, as a key component of economic development, pose great environmental challenges, which underscores the need for close global monitoring to support sustainable urban development. Despite this importance, global city-level maps of industrial land use, especially over multiple years, have been lacking. Here, we present a 10-m resolution global dataset tracking industrial land use in 1,093 large cities (area 100 km² or more) from 2017 to 2023. Using multisource geospatial data and machine learning, the dataset achieves a high overall accuracy of 91.87% to 92.21% across the seven-year period, aligning well with official city maps. We further validated its reliability by computing industrial land area per capita for 1,093 cities, which correlated strongly with per capita CO