Based on Landsat series satellites, China GF-1, GF-6 and other satellite images, GFC30_2020 achieved a spatial resolution of 30 m, using machine learning algorithms with global ecological geographic partitioning and crowd source sample data. The data source uses the 2020 global forest vegetation growth season images, and when the data quality cannot meet the requirements due to cloud coverage or other reasons, data with similar time will be selected. The overall accuracy of the product is higher than 85%.