Abstract: Phosphorus (P) in red paddy soil is easily fixed by iron and aluminum, resulting in low P availability. Excessive phosphate fertilizer applications increase the risk of P leaching. The soil productivity, P pool balance, and characteristics of the P fractions at different levels of phosphate supply were studied in order to provide a basis for the efficient management of P in paddy fields across southern China. The study was based on a 30-year experiment in Fujian yellow-mud paddy fields. The experiment consisted of three P supply treatments in each season: control with no phosphate fertilizer (CK), 30 kg·hm^-2 P₂O₅ (P1), and 60 kg·hm^-2 P₂O₅ (P2). Double-cropping rice was planted from 1987 to 2004, and single-cropping rice was planted from 2005 onwards. The rice yields and P fractions changes under different P supply treatments over 30 consecutive years were studied. The results showed that the average grain yields of the early-season rice, late-season rice, and single-cropping rice significantly increased by 64.9%, 37.0%, and 19.9% in P1 treatment, and 67.0%, 41.2%, and 20.4% in P2 treatment, respectively. Early rice showed the highest increase in yield, followed by late rice and single-cropping rice. Under P2 treatments, in the 31^st year, the soil available P and total P contents in P2 treatment significantly increased by 190.5% and 32.4%, respectively, compared to P1 treatment. The Al-P, Fe-P, Ca-P, and total inorganic P contents also significantly increased, and the Al-P, and Fe-P fraction proportions significantly improved by 2.12 and 4.40 percentage points, respectively. However, the O-P fraction proportion significantly decreased by 9.45 percent points. Phosphate fertilizer applications generally increased the labile organic P (LOP) and medium labile organic phosphorus (MLOP) contents, but decreased the highly stable organic phosphorus (HSOP) content, especially in P2 treatment. Increasing the phosphate fertilizer application rate led to a rise in the proportion of LOP and MLOP to organic P, but it reduced the proportion of HSOP to organic P. Rice grain or straw yields were significantly positively correlated with Al-P, Fe-P, Ca-P, MLOP, and LOP contents. In conclusion, continuous applications of phosphate fertilizer significantly increased the grain yield in yellow-mud paddy soil over 30 consecutive years, but there was no significant difference in the grain yield between P1 and P2 treatments. However, apparent P deficiency was observed in P1 treatment. Increasing phosphate application rate led to a rise in the Al-P, Fe-P, and Ca-P proportions. The results showed that organic P fractions with lower activity changed into fractions with higher activity as the phosphate fertilizer application rate increased. The application of 60 kg(P₂O₅)·hm^-2 phosphate fertilizer per cropping season achieved the optimum apparent balance among P nutrients and maintained an appropriate level of available P.