As the world’s largest cement producer, China urgently requires accurate and comprehensive datasets to support environmental policy development. However, existing datasets are outdated, limited in scope, and lack detailed information. In this study, t-cement plant dataset in China was constructed using an improved Yolov5-IEG model and integrating multi-source data to addresses these gaps. First, an improved Yolov5-IEG t-cement plants detection model using high-resolution remote sensing images was constructed to determine the areas and locations of t-cement plants, resulting in a comprehensive dataset of 953 records. Then, 23 detailed attributes including 11 production status information of each t-cement plant were extracted based on long-term ACF products from 2012 to 2022, Point Of Interest (POI) data, and corporate enterprise information.