The amount of phosphorus (P) dissolved in a closed‐incubation system, in soils receiving Christmas Island grade‐A phosphate rock (CIPR), Gafsa phosphate rock (GPR), and triple superphosphate (TSP), as measured by extraction with 0.5M NaHCO3(APb) or 0.5M NaOH (AP) and expressed as ?Pb/?P*100 (PDP) was compared to P uptake (?Ps) bySetariain a glasshouse experiment. There was no direct relationship between APs and PDP for CIPR, GPR, and TSP added at 50 and 150 mgP/kg soil to three Malaysian soils (Bungor, Kundor, and Segamat) during a 10‐month period. Averaged across soils, rate of addition, and P sources, ?Ps was 17% higher than PDP. Overestimation of PDP by ?Ps could be due to the ability of the roots ofSetariato use more of the P which is dissolved from the three P sources and then chemisorbed, than can be extracted by the NaHCO3reagent. The chemisorbed P pool is extractable using 0.5M NaOH. There was a close relationship between ?Ps and ?P, with a correlation coefficient of 0.85**. Residual P, determined by Pb method after each harvest, successfully predicted Ps bySetariain the subsequent harvest with correlation coefficients varying between 0.74* and 0.99** for CIPR, GPR, and TSP in five soils (Bungor, Durian, Kundor, Segamat, and Tok Yong). The critical Pb values ranged from 5 for Durian to 10 mgP/kg soil for Kundor.