Context. In Sweden, mixtures of 80-85% ground limestone and 15-20% slaked lime (hereafter, 'structure lime') are used in subsidised environmental schemes to improve aggregate stability and mitigate phosphorus losses on clay soils. Aims. This study investigated different rates of structure lime application and soil variables on aggregate stability on clay soils, and whether soil properties can predict aggregate stability following structure liming. Methods. Increasing application rates of 0-16 t ha −1 of structure lime (SL0, SL4, SL8 and SL16) were tested in 30 field trials in Sweden. Soil aggregates (2-5 mm) were collected 1 year after liming and subjected to two rainfall events in a rain simulator. Key results. Leachate turbidity after the second simulated rainfall event decreased significantly (13% and 20%, respectively, in SL8 and SL16) compared with SL0, indicating improved aggregate stability. There was a near-significant interaction (P = 0.056) between treatment and trial. Grouping by initial pH ðH 2 OÞ (range 6.2-8.3), clay content (10-61%), soil organic matter content (SOM, 2.2-7.1) and clay mineralogy (SmV index, 0.2-3.8) revealed different effects on turbidity. Discriminant analysis of soil characteristics and four tillage variables correctly classified the outcome for 27 of the 30 trial sites. Conclusions. Results show that structure liming can improve aggregate stability 1 year after liming, and can thereby prevent particulate P losses from soils with high clay and SOM content, low SmV index and low initial pH. The discriminant analysis also showed the importance of tillage for the outcome of structure liming. Implications. Clay soil characteristics such as SOM and pH significantly affected aggregrate stability after structure liming.