This work describes a metabolic profiling study of gilthead sea bream, from three different aquaculture systems, using H-1 NMR and chemometrics. A total of 54 samples under two different storage regimens were analysed. The assignment of all major NMR signals of the perchloric extracts was performed. A comprehensive multivariate data analysis proved able to distinguish the fish metabolism amongst the different aquaculture systems and to determine whether a fish was stored or not. The state of energy metabolism of inosine proved a robust biomarker for evaluating storage time. A new multivariate classification tool, iECVA, revealed several metabolites which are important biomarkers for characterising the three different aquaculture systems: glycogen (stress indicator), histidine, alanine and especially glycine for long storage times and mainly betaine for fresh samples. The findings represent a step forward in understanding how in vivo and postmortem processes affect the total quality of the final product