Models that estimate rates of energy flow in complex food webs often fail to account for species-specific prey selectivity of diverse consumer guilds. While DNA metabarcoding is increasingly used for dietary studies, methodological biases have limited its application for food web modeling. Here, we used data from dietary metabarcoding studies of zooplankton to calculate prey selectivity indices and assess energy fluxes in a pelagic resource-consumer network. We show that food web dynamics are influenced by prey selectivity and temporal match-mismatch in growth cycles and that cyanobacteria are the main source of primary production in the investigated coastal pelagic food web. The latter challenges the common assumption that cyanobacteria are not supporting food web productivity, a result that is increasingly relevant as global warming promotes cyanobacteria dominance. While this study provides a method for how DNA metabarcoding can be used to quantify energy fluxes in a marine food web, the approach presented here can easily be extended to other ecosystems.