The oceans are important regulators of the Earth’s climate system by sequestering carbon from the atmosphere taken up by primary producers. Zooplankton, including protozoans and metazoans of different phyla and size classes, occupies several trophic niches and regulates energy flow between primary producers and fish. The structural configuration of the food web determines the rates at which primary production is either enriched to sustain organisms at higher trophic levels or exported to the ocean floor. However, limited knowledge about plankton interactions causes uncertainty of how the oceans will respond to climate changes. This thesis presents a framework for studying and modeling pelagic food webs using novel implementations of DNA metabarcoding. Study I shows that DNA metabarcoding of zooplankton sampled in their natural environment reveals a broader and more complex diet than zooplankton in classic grazing observations. We also show that differential feeding strategies facilitate species coexistence and that the zooplankton diet is largely dependent on prey availability. The approach was extended in Study II, where we include the smaller fraction of zooplankton that is often overseen in food web studies to broaden the perspective of functional diversity in pelagic food webs. We show that different populations have unique functions in channeling the primary production of different sources and especially highlight the role of filter-feeders in making detrital nutrients available for other organisms in the food web. In Study III, we shifted focus to trophic links between zooplankton and fish by comparing niche overlap between the three main planktivorous fish in the Baltic Sea - stickleback, sprat, and herring. The results from the three first studies were finally used to calculate selectivity indices between each predator and prey. This information was implemented in Study IV in a network model quantifying fluxes of energy through the food web. The model revealed cyanobacteria as the primary contributor to secondary production in the Baltic Sea food web and that the spring bloom of diatoms and dinoflagellates remains largely unutilized by the zooplankton. This is the first time DNA metabarcoding is used to compare niche differences of several zooplankton species in a pelagic guild and to quantify fluxes in a food web model. The thesis refines our knowledge of pelagic community and food web structure, and the framework presented here is a suitable entry point for food web modeling in other ecosystems.