The purpose of this study was to design an in vitro experiment that can assess the stability of polymeric micellar formulations of hydrophobic drugs such as cyclosporine A (CyA) in blood, and predict the in vivo performance of the examined delivery system. Poly(ethylene oxide)-block-poly(epsilon-caprolactone) (PEO-b-PCL) copolymers were assembled to polymeric nano-containers for the physical encapsulation of CyA by a co-solvent evaporation method using different loading conditions. CyA-loaded micelles were prepared and compared to commercially available intravenous formulation of CyA (Sandimmune) for in vitro release, protein binding, and pharmacokinetic parameters in Sprague-Dawley rats. The unbound fraction (fu) of CyA was determined using an erythrocyte vs. plasma and buffer partitioning technique. Different polymeric micellar formulations of CyA did not show any significant difference in CyA release when dialyzed against bovine serum albumin. The fu experiments, however, revealed a significant decrease in the fu of the loaded drug with an increase in the drug/polymer loading ratio, while the fu of all micellar formulations were significantly lower than Sandimmune. The pharmacokinetic study showed that fu of CyA in each formulation correlated with its in vivo performance determined by pharmacokinetic parameters: the lower fu of the formulation, translated to a higher area under the concentration versus time curve (AUC), and a lower clearance (CL) and volume of distribution (Vd). In conclusion, determination of the unbound fraction of encapsulated drug can be used to predict the in vivo stability of polymeric micellar nano-containers. PEO-b-PCL micelles containing higher CyA-loaded levels are shown to be more stable changing the pharmacokinetics of the encapsulated CyA to a higher extent.