Fish, even of the same species, can exhibit substantial variation in energy density (energy per unit wet weight). Most of this variation is due to differences in the amount of storage lipids. In addition to their importance as energy reserves for reproduction and for survival during unfavourable conditions, the accumulation of lipids represents a large energetic flux for many species, so figuring out how this energy flux is integrated with other major energy fluxes (growth, reproduction) is critical for any general theory of organismal energetics. Here, we synthesize data from a wide range of fish species and identify patterns of intraspecific variation in energy storage, and use these patterns to formulate a general model of energy allocation between growth, lipid storage and reproduction in fishes. From the compiled data we identified two patterns: (1) energy density increases with body size during the juvenile period, but is invariant with body size within the adult size range for most species, and (2) energy density changes across seasons, with depletion over winter, but increases fastest in periods of transition between favourable and unfavourable conditions for growth (i.e. fall). Based on these patterns we propose DEBlipid, a simple, general model of energy allocation that is closely related to a simplified version of Dynamic Energy Budget theory, DEBkiss. The crux of the model is that assimilated energy is partitioned, with κ fraction of energy allocated to pay maintenance costs first, and the surplus allocated to growth, and 1 - κ fraction of assimilated energy is allocated to accumulating storage lipids during the juvenile phase, and later to reproduction as adults. This mechanism, in addition to capturing the two patterns that motivated the model, was able to predict lipid dynamics in a novel context, the migration of anadromous fish from low-food freshwater to high-food marine environments. Furthermore, the model was used to explain intra and interspecific variation in reproductive output based on patterns of lipid accumulation as juveniles. Our results suggest that many seemingly complex, adaptive energy allocation strategies in response to ontogeny, seasonality and habitat quality can emerge from a simple physiological heuristic.