Existing models of marine ecosystems address specific issues related to the bottom-up forcing of production or to the top-down effects of fishing on a limited range of the trophic spectrum. Very few existing models explicitly incorporate the dynamics from one end of the ecosystem to the other and thus allowing the exploration of interplay between exploitation and climate effects. The shift to an ecosystem approach to fisheries and concerns about the ecological effects of climate change require the assemblage of knowledge assembled from the respective marine disciplines with the view to build end-to-end models of marine ecosystems. Here, with a focus on plankton and fish models, we present some issues and recommendations for the integration of models between trophic levels (vertical integration) and within functional groups (horizontal integration within trophic levels). At present, vertical coupling of plankton and fish models is mainly realized through predation processes, generally represented as a functional response. In the absence of empirical evidence and quantification, the choice of the functional response term is often made by default, and is reduced to a parameterization problem. A strategy is proposed to overcome this arbitrary choice. In addition to the vertical coupling of trophic models, the structure of end-to-end models incorporates biodiversity via horizontal integration of trophic levels. For guiding the selection of key components to be included in end-to-end models, the idea that marine food webs are structured as alternative trophic pathways is highlighted and related to observed dynamics. We suggest that an important early step in model development is the identification of major trophic pathways and bottlenecks in an ecosystem using a historical perspective.