How grazer selectivity regulates the primary producer community is a core topic in ecology. Yet, the role of zooplankton grazing selection on phytoplankton dynamics is poorly understood. Few studies have compared the effect of grazers with contrasting selectivity on mixed phytoplankton prey, and none over multiple phytoplankton generations. We tested the hypothesis that a selectively grazing copepod (Eudiaptomus gracilis) would facilitate the dominance of a toxic cyanobacterium (Microcystis aeruginosa) by grazing on a competing eukaryotic microalga (Cryptomonas pyrenoidifera), while a generalist cladoceran (Daphnia magna) would have no effect on the dominance of cyanobacteria in 4-d laboratory cocultures. Experiments started with a ninefold initial dominance of Cryptomonas over Microcystis by biomass. Each grazer type was added to cocultured phytoplankton and the abundance of phytoplankton was compared to no-grazer controls. As predicted, Daphnia had no effect on the relative abundance of its prey and the copepod facilitated Microcystis dominance, although the strength of facilitation slightly declined with time. As the copepod reduced mostly the biomass of the edible algae, it pushed the system toward the dominance of toxic prey, which likely reduced the efficiency of selective grazing on the last day. Hence, while the selective grazer promoted cyanobacterial dominance, the effect may be weaker than predicted from extrapolating grazing rates obtained from short-term (i.e., hourly) assays. Overall, predicting the role of zooplankton selectivity on phytoplankton dynamics—especially harmful algal blooms—would benefit from accounting for fluctuations in grazer effects due to shifting abundance and growth of each prey over time.