Abstract: To evaluate the possible ecological risks of transgenic (CAgcGH) common carp, an experimental man-made lake was set up in 2002. We investigated the phytoplankton community structure, seasonal dynamics, annual variations of phytoplankton, as well as its relationships with environmental factors. Samples of phytoplankton were collected seasonally from 2006 to 2010. Sixty-six species belonging to seven phyla and forty-seven genera were identified, among which Chlorophyta was the most. The results of two-way indicators species analysis (TWINSPAN) and detrended correspondence analysis (DCA) showed that the samples were congregated into four groups, which indicated obvious seasonal character. In winter, the community structure was simple.The biodiversity was the lowest. The community was mainly composed of Cyclotella sp. and Dinobryon divergens; in spring, several diatoms, such as Cyclotella sp., Synedra sp. and Melosira granulata became dominant species; in summer, community structure was the most complex, with the highest biodiversity, and Merismopedia glauca and Lyngbya contarta were dominant. In our five-year-experiment, the cell density rose by 33.1%, and mean annual cell density was (1.430.75) 106 cells/L; the percentage of diatoms in the phytoplankton community dropped from 48.2% to 16.2%, while the percentage of Cyanobacteria in the phytoplankton community rose from 9.3% to 42.2%. The results of canonical correlation analysis (CCA) demonstrated that the community structure was mainly influenced by temperature and dissolved oxygen. The influence of the density of total phosphorus and nitrogen was non-negligitible, and the influence of pH was small. Diatoms mainly appeared in samples from middle and low temperature sites, blue-green algae mainly appeared in samples from high temperature sites, Desmidiales mainly appeared in samples from sites with high transparency and total phosphorus, and Chrysomonadales mainly appeared in samples from sites with high dissolved oxygen and low temperature.