Surface soils interact strongly with both climate and biota and provide fundamental ecosystem services that maintain food, climate and human security. However, the quantitative linkages between soil properties, climate and biota remain unclear at the global scale. By compiling a comprehensive global soil database, we mapped eight major soil properties (bulk density; clay, silt, and sand fractions; soil pH; soil organic carbon, SOC, density; soil total nitrogen, STN, density; and soil C : N mass ratios) in the surface soil layer (0–30 cm), based on machine learning algorithms, and demonstrated the quantitative linkages between surface soil properties, climate and biota at the global scale, which we call the global soil–climate–biome diagram. In the diagram, bulk density increased significantly with higher mean annual temperature (MAT) and lower mean annual precipitation (MAP); soil clay fraction increased significantly with higher MAT and MAP; soil pH decreased with higher MAP and lower MAT and the “critical MAP”, which means the corresponding MAP at a soil pH of = 7.0 (a shift from alkaline to acidic soil), decreased with lower MAT. SOC density and STN density were both jointly affected by MAT and MAP, showing an increase at lower MAT and a saturation towards higher MAP. Surface soil physical and chemical properties also showed remarkable variation across biomes. The soil–climate–biome diagram suggests shifts in soil properties under global climate and land cover change.