Several proteomics studies have been conducted to identify new cerebrospinal fluid (CSF) biomarkers in Multiple Sclerosis (MuS). However, the complexity of CSF and its invasive collection, limits its use. Therefore, the goal of biomarker research in MuS is to identify novel distinctive targets in CSF or in easily accessible biofluids. Tears represent an interesting matrix for this purpose, because (1) tears are related to the central nervous system (CNS) and (2) the CNS contains Extracellular Vesicles (EVs) derived from brain cells. These EVs are emerging new biomarkers associated to several neurological disorders. Here we applied an optimized flow cytometer for the identification and subtyping of EVs from CSF and tears. We found, for the first time, microglia-derived and neural-derived EVs in tears. The flow cytometer was used to sort and purify 106 EVs from untouched CSF and tears of MuS patients and healthy subjects. Purified EVs were analyzed with shotgun proteomics analysis, revealing that EVs from both CSF and tears of MuS patients conveyed similar proteins. Our data demonstrated a specific EVs-mediated molecular cross talk between CSF and tears, which opens the door to new diagnostic perspectives for MuS.