The stable isotopes of Phenylalanine (Phe) and Tyrosine (Tyr) are often used to study whole body protein metabolism in humans. Non-compartmental approaches give limited physiological insight in the compartmental characteristics. We therefore developed a mathematical model of Phe/Tyr metabolism to describe protein fluxes using stable tracer dynamic data in plasma following iv bolus of L-[ring-¹³C₆]-Phe and L-[ring-²H₄]-Tyr in healthy subjects. The model consists of four compartments describing Phe/Tyr kinetics. Since the model is a priori non-identifiable, it is quantified in terms of two uniquely identifiable submodels representing two limit case scenarios, based on known physiology. The two submodels, identified by using the software SAAM II, fit well the experimental data of all individuals and provide an unbiased overview of the metabolic pathway in terms of intervals of validity of the non-uniquely identifiable variables. The model provides estimates of the flux from Phe to Tyr (4.1±1.0 µmol•kg ffm^-1•h^-1, mean±SEM) and intervals of validity of the flux and pool estimates. Our preferred submodel yielded protein breakdown flux (50.5 ± 5.2µmol•kg ffm^-1•h^-1), net protein breakdown (4.1±1.0µmol•kg ffm-1•h^-1), Tyr from Phe hydroxylation (~12%), hydroxylated Phe (~8%) and flux ratio of Tyr/Phe arising from protein catabolism (0.68), consistent with available literature. The other submodel suggest that the assumptions made by non-compartmental analysis are consistently underestimated. Our accurate and detailed model for estimating Phe/Tyr metabolic pathways in humans might be essential in applications in a variety of scenarios describing whole body protein synthesis and breakdown in health and disease.