Most of P on peat is in the form of organic P. Composition of P on peat, especially organic P, is rarely studied. Therefore fractionation of P on peat is important to know the proportion of inorganic and organic P and available P. The aim of the experiment was to study the composition of P added with mineral soil ameliorant and P fertilization. The experiment was conducted in the laboratory of the Indonesian Center for Soil and Agroclimate Research and Development, Bogor. Two hundreds g peat dry weight (105o) added with mineral soil ameliorant in the rate of 7.5% maximum Fe sorption which is equal to 10.4 g pot-1. The mixture of peat and mineral soil (Oxisols) was then added with P fertilizer and followed by incubation for 4 weeks. After incubation, the mixture was fractioned for P. P treatment consisting of four P fertilizer sources that are Maroko rockphosphate, Christmas rock phosphate, Ciamis rock phosphate and SP-36. Four P levels are 25, 50, 75, and 100% of P sorption plus partial complete control, and replicated two times. Composition of P in this study that is readily labile, moderately labile, and non labile. P fraction determinated with sequential extraction based on the method had been used by Ivanoff. The results showed that peat from Air Sugihan Kiri, South Sumatrawas poor (oligotropyc), acidic, and low exchangeable K and ash content. Most of P on the peat was in the proportion of organic P (around 77 up to 95%) and the rest was in a small amount, inorganic P (10 up to 23%). The composition of P organic fraction showed that the non labile was relatively the same P proportion with moderately labile, meanwhile the readily labile was low. Source and rock phosphate level did not affect theproportion of organic P. The application of Maroko and Ciamis rock phosphates at the level of 100% of P sorption increased available P (readily P inorganic). Non-labile inorganic P proportion from Christmas rock phosphate was higher than those from Maroko and Ciamis.