Abstract The distribution and form of P in soil is central to the sustainability of agricultural practice. This study used sequential fractionation and 31P nuclear magnetic resonance spectroscopy ( 31P NMR) of NaOH–EDTA extracts to examine the influence of pastoral, native (undisturbed) and forest land use on soil P forms in 5 contrasting soils ranging from a Regosol to a Rendzina in Otago, New Zealand. Climatic factors likely to influence soil P distribution were negated by careful site selection. Together with a decrease in soil organic C (31%), total P decreased in forested soils (mean=674 mg kg −1) compared to native soils (mean=784 mg kg −1). In contrast, the ratio of inorganic to organic P increased (10%) probably due to mineralization of organic P in forest soils, while for pasture soils, accumulation of P in inorganic forms due to P inputs via fertilisers and animal dung was to blame. Investigation of the organic P forms in NaOH–EDTA extracts of each land use by 31P NMR indicated that diesters were greatest in the native soil (4–12% of total P in spectra), and declined as a proportion of total P in pasture soils and more so in forest soils. This was reflected in a decline of the diester to monoester ratio. However, the ratio was generally greater in forest than pasture soils and attributed to the labile nature of diesters, mineralization of monoesters in forest soils, and an increase in monoesters in pasture soils from inositol phosphates in plant debris. This effect was pronounced in the Regosol due to sandy texture and the preferential accumulation of plant debris in coarse particle size fractions. Due to the depletion of soil P reserves, forest soils in the area should be followed by pasture and well managed fertiliser additions before replanting.