Phosphorus (P) losses from many agricultural soils are driven by fast transport processes, such as surface runoff, shallow interflow and macropore flow. Close interaction of rain with topsoil layers (0–5 cm), rich in available P, results in high P concentrations in these rapid runoff pathways. This is particularly pronounced in permanent grassland soils typically characterized by a P enrichment of the topsoil. One option to mitigate P losses from grassland is therefore to reduce soil P content, particularly available P, in this critical surface layer. Omitting application of P fertilizer will reduce soil P in the long term but more drastic methods may be necessary to achieve P loss reductions in shorter timescales. In this study, three management options to reduce P in runoff from grassland soils were evaluated in a field experiment lasting 2 years on two grassland sites: (i) no P inputs, (ii) no P inputs plus soil tillage and (iii) no P inputs plus tillage combined with application of a P sorbing soil amendment (iron oxide). Withholding P inputs did not decrease either the available P in the soil or the P concentrations in runoff from sprinkler experiments over a 2-year period. Both the treatments which included tillage caused immediate reductions in both soil available P and runoff P, with Fe ⁄ OH amendment being more effective. However, in the second year the effects of both tillage treatments were less apparent and only the Fe ⁄ OH treatment significantly decreased runoff P. The results indicate that the long-term enrichment of soils due to P application in excess of plant needs is unlikely to be remedied by short-term treatments. However, further work is required to assess the applicability of our conclusion to other sites.