Three gorges reservoir (TGR), the largest water conservancy project ever built in the world, with about 348.93 km² of littoral zone produced due to the water level control, attracted more and more attention for its fragile environment. The variation of plant community and nutrient-release of the soil in the littoral zone, which were caused by long-term winter flooding and summer drought, are of great concerns for the ecological security of the TGR. However, little is known about the relationship between plant community and the combined forms of the soil nutrients. In this study, Pengxi river, an typical tributary of the TGR, with the largest area for littoral zone, were selected to perform an investigation on the phosphorus forms of the rhizosphere and non-rhizosphere soils of four typical plants (contained Cynodon dactylon L., Cyperus rotundus L., Anthium sibiricum Patrin ex Widder, and Zea mays L.) from three transects with different soil types (Paddy soil, Purple soil, and Fluvo-aquic soils). The variables measured included total phosphorus (TP), available P (AP), inorganic phosphorus (IP), organic phosphorus (OP), and different forms of IP and OP (i.e weakly adsorbed phosphorus WA-P, potential active phosphorus PA-P, Fe/Al-bound phosphorus Fe/Al-P, Calcium bonded phosphorus Ca-P and Residue phosphorus Re-P). The results showed that soil type was a crucial factor of determining nutritional status and species of phosphorus, while vegetation cover was an additional factor. 1) Compared with all non-rhizosphere soil, concentrations of TP, AP, IP, OP, and different phosphorus forms were obviously enriched in rhizosphere soil of different plants in three soil types as the result of the root absorption and activation. The TP, AP, IP and OP concentrations showed C. dactylon > A. sibiricum > C. rotundus > Z. mays. Enrichment rate of the IP showed the highest in the rhizosphere of C. dactylon and the lowest in Z. mays. 2) WA-P_i, PA-P_i and Fe/Al-P_i (representing some labile phosphorus) in rhizosphere of A. sibiricum and Z. mays were higher than those in the rhizosphere of C.rotundus and C.dactylon. The Ca-P_i and R-P_i, which are mainly stabilized phosphorus, were much lower in rhizosphere of A. sibiricum and Z.mays than those in C.r otundus and C. dactylon. Those indicated that the rhizosphere effect of different plants on the IP cycle was significant differences. The activation of stabilized phosphorus for the rhizosphere effect of A.sibiricum and Z.mays were much more obvious than that of C.rotundus and C.dactylon. 3) WA-P_o, PA-P_o, Fe/Al-P_o, Ca-P_o, R-P_o (different OP forms) were also showed a similar law with IP forms. WA-P_o and PA-P_o in the rhizosphere of C. rotundus and C. dactylon were much lower than those in rhizosphere of A. sibiricum and Z. maysL, while Fe/Al-P_o were lower in rhizosphere of A. sibiricum and Z. mays. Higher enrichment effect of WA-P_o and PA-P_o under Z. mays zone showed the higher activation efficiency of phosphorus in rhizosphere of Z. mays. 4) pH, organic matter and volume-weight was significantly correlated with the IP forms, while only pH related with OP forms. Our research could provide some indirect evidences that C. rotundus and C. dactylon was better than A. sibiricum and Z. mays to improve soil P holding and fixation. The vegetation recovery in the drawdown zone should consider the rhizosphere effect of different plants on N cycle.