Samples of soil and plants were collected from shoal wetlands in the Poyang Lake for indoor simulation of static transfer of phosphorus in a soil-water-plant system to explore laws of the absorption, release and transfer of phosphorus in the soil-water-plant wetland system of the Poyang Lake. Before and after the experiment, transfer of phosphorus relative to form in the system was analyzed. The experiment device was a cylindrical container of polyvinyl chloride, 10cm in inner diameter and 40cm in height, containing a 10cm thick layer of homogeneous sediment. Water lily was planted in the cylinder. The overlaying water in the cylinder was sampled for and analyzed of phosphorus on D3, D6, D9, D16, D23, D28, D35 and D42 after the experiment began (D day). Meanwhile, soil and plant samples were also collected for analysis of phosphorus concentration. At the end of the experiment, fractionation of phosphorus in the surface layer (1～2cm) was performed, and concentrations of Fe-P, Ca-P and Al-P were determined to explore variation of the transfer of phosphorus relative to form. On such a basis, the Freundlich model and Langmuir model were used to perform one-dimension linear regressive analysis of adsorption and desorption of P in the soil sediment of the shoal wetlands in the Poyang Lake. Results show that phosphorus was released continuously from the soil into the water, and mostly absorbed by plants. Before the experiment, phosphorus concentration was found to be 1.0 mg g^-1 in the plants and 2.3 μg mL^-1 in the overlaying-water, and after the experiment it was 2.1 mg g^-1 and 0.062 mg L^-1, respectively. In the initial period of the experiment, phosphorus concentration in the overlaying water decreased from 1.7 to 1.4 mg L^-1, and on D16, it was found to be 0.5 mg L^-1, and then it leveled off from D35 and on. During the first 3 days, phosphorus in the soil decreased quickly, from 25.5 to 11.5 mg kg^-1. As this was only a static experiment, though part of the phosphorus in the soil was released into the overlaying water, the exchange between the soil and the water was very slow. P concentration in the overlaying water as a whole still displayed a decreasing trend. Only the concentration in the water layer close to the surface of the soil increased somewhat. At that time, phosphorus content in the plants kept rising rapidly. Plant roots grew deep into the soil, facilitating phosphorus transfer from the soil into the plants. With the growth of the plant root systems, plants’ demand for P increased. The rising P content in the plant began to level off on D28. The decline of phosphorus content in the overlaying water began to slow down at the mid-later stages of the experiment. At that time, the plants were well developed and phosphorus content in the soil kept almost unchanged, and the phosphorus in the soil-water-plant system reached its equilibrium. Meanwhile, soil phosphorus fractionations before and after the experiment show that the content of Al-P remained almost unchanged, the content of Ca-P increased slightly but the content of Fe-P did by a large margin. The content of Fe-P in the soil was quite high and that of Ca-P and Al-P quite low, either before or after the experiment. Among the three fractions, Al-P was extremely low in content, being 0.020 mg kg^-1 before the experiment and 0.016 mg kg^-1 after the experiment, almost unchanged; the content of Ca-P increased slightly from 0.202 to 0.239 mg kg^-1; and the content of Fe-P increased by a large margin from 0.746 to 0.862 mg kg^-1. Relative to the Freundlich model, the Langmuir model was better in simulating phosphorus adsorption process in wetland soils. The findings of this study may serve as scientific theoretical basis for management of lake pollution and research on environmental effects of pollutant in wetlands.