Cascade reservoir-riversystems evolve to becomea new ecological region produced by anthropogenic disturbances. Systematic studies of spatial-temporal patterns of phytoplankton community structure are still lacking in these systems, and thus we investigated phytoplankton composition and related environmental factors in the impounded Wujiang River, southwest China. The results showed that the dominant groups were Chlorophyta, Bacillariophyta, and Cyanophyta, whereas Pyrrhophyta and Cryptophyta mainly appeared from May to August. Bacillariophyta were dominant algae in the river water, Cyanophyta in the eutrophic reservoirs, and Chlorophyta, Bacillariophyta, and Cyanophyta were dominant in the mesotrophic reservoirs. Water temperature was an important limiting factor for algal growth, and algal biomass produced 2 peak values in May and in August. Backpropagation artificial neural networks (BPANNs) have accurately predicted time-series variation of dominant algal abundance. The sensitivity analyses indicated that water temperature, dissolved CO2 and silicon were the main factors influencing phytoplankton community structure, suggesting algal succession is constrained by carbon and silicon biogeochemical cycles. Results suggested that Cyanophyta used HCO3- as an inorganic carbon source when the CO2 concentration was less than similar to 10 mu mol L-1, leading to the species succession from Bacillariophyta to Cyanophyta.