Energy content has long been proposed as a fundamental, integrated, and reliable indicator of the condition of individuals as it reflects past bioenergetics and influences future life‐history traits. There is a direct biochemical link between energy density and body composition described by four main compounds in fish (protein, lipid, ash, and water), with proteins and lipids being the sources of energy. If relationships between water content, or lipid content, and energy density have been well described in relative terms, the absolute mass variations in the proximate composition have been overlooked and thus their interpretation is often equivocal. In our study, based on a large and unique dataset on the proximate composition and energy density of anchovy (Engraulis encrasicolus) and sardine (Sardina pilchardus) from sampling in the Bay of Biscay and the English Channel, we aimed to better explain the patterns between water content and other proximate components or energy density, based on the dynamics of proteins, lipids, and water absolute masses. For the first time, we defined good, intermediate, and poor condition states in wild fish, based on water content, corresponding to the different dynamics of lipids and proteins in the metabolism of individuals. Anchovy and sardine exhibited remarkably similar patterns of variation in the compounds and in the limits between the condition states with respect to water content. Those patterns revealed that water mass remained constant for a given fish size whatever its condition state, and that variability in water content only resulted from the variation in lipid and protein masses. Furthermore, the differential dynamics of proteins and lipids, with only lipids mobilized in the good condition state, only proteins in the poor condition state, and both proteins and lipids in the intermediate condition state, elucidates the nonlinear pattern observed in the relationship between energy density and water content. Overall, our results highlight the importance of monitoring the intraspecific variations in water content to predict the proximate composition and energy content in small pelagic fish and better assess individual and population conditions in changing ecosystems as well as to better parameterize bioenergetic models.