The effects of various management practices on soil phosphorus (P) dynamics were investigated in a field experiment in New South Wales, Australia, during 24 years of different crop rotation, stubble management, and tillage treatments. Topsoil samples collected at the beginning of the trial and after 6, 12, 18, and 24 years were analysed for resin-extractable P, inorganic and organic P, and total P. According to the calculated P input–output budget, 9–14 of the 20 kg P/ha added as superphosphate annually remained in the system, depending on the treatment. The measured increase in total P in 0–0.20 m did not differ between treatments, showing an accumulation rate of only 9 ± 2 kg P/ha.year. These results suggest a loss of 4 ± 2 kg P/ha.year, presumably into lower soil layers. Resin-extractable P at 0–0.10 m increased by 1.7 kg P/ha.year, irrespective of the treatment. The increase in total P after 24 years was almost completely accounted for by the increase in total extractable inorganic P. Changes in organic P paralleled changes in organic carbon, with a significant loss in treatments with stubble burning (wheat–lupin rotation and continuous wheat), and a significant accumulation in a wheat–subterranean clover rotation with stubble retention and direct drilling. We conclude that on the time scale of this experiment, the dynamics of carbon and organic P are closely linked.