Predicting changes of phytoplankton communities in response to global warming is one of the challenges of ecological forecasting. One of the constraints is the paucity of general principles applicable to community ecology. Based on a synecological analysis of a decadal-scale database in China Seas, we created a niche habitat classification scheme relating nine phytoplankton groups to fifteen statistically refined realized niches comprised of three niche dimensions: temperature, irradiance, and nitrate concentrations. The niche scheme assigned the nine phytoplankton groups to three types of niches: a cold type, a warm type, and a type associated with high irradiance and high nitrate concentrations. The fact that phytoplankton groups in cold niches were governed by irradiance and those in warm niches by nitrate is consistent with general ecological theories, but the fact that diatoms were the only dominant group in high-irradiance, high-nitrate niches challenges the idea based on autecological studies that diatoms are generally better adapted to low-irradiance, high-nutrient conditions. When combined with an irradiance model, the niche scheme revealed that photoinhibition of Prochlorococcus, which is predicted from autecological studies, is a function of temperature. We used the niche scheme to predict the responses of phytoplankton communities to environmental changes due to seawater warming and eutrophication. The results of the study suggest that a synecological analysis of large databases from field studies facilitates identification of general principles of community ecology that can be used to forecast responses of biological communities to environmental changes.