Glucose oxidase (GOD) has been encapsulated as a model protein within poly-DL-lactide-poly(ethylene glycol) (PELA) microspheres to evaluate the activity retention during microencapsulation process. This paper was aimed to investigate the effect of process parameters, such as the preparation method, the used matrix polymer with different compositions, the solvent system and the addition of stabilizer on the structural integrity and activity retention of encapsulated protein. The stability of the protein released during in vitro assay was also assessed. The obtained results showed that the solvent extraction/evaporation method based on the formation of double emulsion w/o/w benefited the activity retention 1 2 compared with the phase separation method based on the formation of w/o/o. And in the emulsion-evaporation system 1 2 most of the protein activity was lost during the first emulsification procedure to form primary emulsion w/o(ca. 28%) and 1 the second emulsification procedure to form the double emulsion w/o/w (ca. 20%), in contrast to other processes 1 2 occurring during microspheres preparation. The matrix polymer and the solvent system in the oil phase had an impressive impact on the activity retention, while the addition of gelatin in the internal aqueous phase resulted in no major reduction of activity loss. GOD release from PELA microspheres exhibited a triphasic profile, that is, the initial burst release during the first day, the gradual release over about 1 month, and then the second burst release. The encapsulation of GOD in PELA microspheres was effective in reducing its specific activity loss. Sixty-seven per cent of the initial specific activity retention was detected for the released GOD from microspheres formulation during 1 week of incubation, but nearly all the activity was lost for GOD in solution incubated under the same condition. SDS-PAGE results showed that, although the activity loss was detected, no rough changes of molecular weight of GOD was observed during encapsulation procedure and the initial days of incubation into the in vitro release medium.