Polycystic ovary syndrome (PCOS) is a complex endocrine disorder of heterogeneous etiology. Proteomics techniques have been used for elucidating the physiopathology of PCOS, yet the proteins identified so far were rarely the same across tissues and studies. The present review discusses the current challenges in the application of proteomics to the study of PCOS. A well-defined research design and an appropriate selection of study populations, samples and proteomic platforms are essential in clinical proteomics. Furthermore, the findings derived from proteomic approaches should be validated by complementary techniques, and the reproducibility of the results has ideally to be confirmed by different studies. Only when meeting these requirements, the proteins identified by proteomic techniques should be considered as candidates for future studies aiming to define specific molecular phenotypes of PCOS and their possible role in the metabolic and hormonal abnormalities characteristic of this syndrome.