Background: Progressive multifocal leukoencephalopathy (PML) caused by the JC virus is the main limitation to the use of disease modifying therapies for treatment of multiple sclerosis (MS). Methods: To assess the PML risk in course of ocrelizumab, urine and blood samples were collected from 42 MS patients at baseline (T0), at 6 (T2) and 12 months (T4) from the beginning of therapy. After JCPyV-DNA extraction, a quantitative-PCR (Q-PCR) was performed. Moreover, assessment of JCV-serostatus was obtained and arrangements’ analysis of non-coding control region (NCCR) and of viral capsid protein 1 (VP1) was carried out. Results: Q-PCR revealed JCPyV-DNA in urine at all selected time points, while JCPyV-DNA was detected in plasma at T4. From T0 to T4, JC viral load in urine was detected, increased in two logarithms and, significantly higher, compared to viremia. NCCR from urine was archetypal. Plasmatic NCCR displayed deletion, duplication, and point mutations. VP1 showed the S269F substitution involving the receptor-binding region. Anti-JCV index and IgM titer were found to statistically decrease during ocrelizumab treatment. Conclusions: Ocrelizumab in JCPyV-DNA positive patients is safe and did not determine PML cases. Combined monitoring of ocrelizumab’s effects on JCPyV pathogenicity and on host immunity might offer a complete insight towards predicting PML risk.