Common cropping soils in south-western Australia are Tenosols or deep sands and Kandosols or acid yellow sandy earths and duplex sandy gravels. The region has a Mediterranean climate with a large proportion of the rainfall occurring over June-July that can leach anions from soil. Farmers typically apply 1-10 kg S/ha, mainly at seeding. Sandy topsoils and some Tenosol subsoils typically have low sulfate sorption capacity and under the prevailing leaching environments can contain low levels of sulfate. Seedlings of cereal and canola crops grown in these soils are often deficient in sulfur. Kandosol subsoils have higher sulfate sorption capacity, and hence the ability to retain sulfate fertiliser leached from the topsoil. Crops can subsequently use this sorbed S. In these situations there is often a limited grain yield response to applied S fertiliser. Soil testing is generally done on the top 0.1 m of soil. Such tests fail to account for sulfate in subsoils. To obtain a better sulfate fertiliser recommendation, deeper S testing is required together with a simple method to rank the sulfate sorption capacity of soil. We propose a procedure to estimate the capacity of soil to sorb sulfate that has potential use in commercial soil testing laboratories and in mechanistic models. The modified Freundlich equation was used to describe sulfate sorption by soil. The equation is written as, S = acb – q, where S is the amount of sulfate sorb (g S/kg soil); c is the concentration of sulfate measured in the extract solution (g S/L), and a, b and q are coefficients. Three steps are required to determine the coefficients of the equation. The first step is to measure soil sulfur using the KCl40-S procedure to determine the value of q. The second step is to measure sulfate sorption by soil and the concentration of sulfate in the extract solution. This was done by mixing 3 g of soil with 30 ml of 0.01 M CaCl2 to which of 100 g S/kg soil was added and shaken for 48 hours. Sulfate sorption was calculated by subtracting the amount of sulfate measured in the extract solution from the 100 g S/kg soil added to the suspension. The a parameter was then calculated from the measurements using the formula, a = S + q/cb. The final step was to estimate b, which was determined from a regression equation, fitted against measurable soil properties. The advantage of this approach is the method defines the sulfate sorption curve. Hence, the approach can be applied to mechanistic models. In contrast, pervious single point measurements only define a point on the sorption curves.