Osteoarthritis (OA) is the most common degenerative joint disease. Mesenchymal stromal cells (MSC) are promising cell-based therapy for OA. However, there is still a need for additional randomized, dose-dependent studies to determine the optimal dose and tissue source of MSC for improved clinical outcomes. Here, we performed a dose-dependant evaluation of umbilical cord (UC)-derived MSC (Celllistem) in a murine model and in knee OA patients. For the preclinical study, a classical dose (200.000 cells) and a lower dose (50.000 cells) of Cellistem were intra-articularly injected into the mice knee joints. The results showed a dose efficacy response effect of Cellistem associated with a decreased inflammatory and degenerative response according to the Pritzker OARSI score. Following the same approach, the dose-escalation phase I clinical trial design included 3 sequential cohorts: low-dose group (2 × 106 cells), medium-dose group (20 × 106), and high-dose group (80 × 106). All the doses were safe, and no serious adverse events were reported. Nonetheless, 100% of the patients injected with the high-dose experienced injection-related swelling in the knee joint. According to WOMAC total outcomes, patients treated with all doses reported significant improvements in pain and function compared with baseline after 3 and 6 months. However, the improvements were higher in patients treated with both medium and low dose as compared to high dose. Therefore, our data demonstrate that the intra-articular injection of different doses of Cellistem is both safe and efficient, making it an interesting therapeutic alternative to treat mild and symptomatic knee OA patients. Trial registration ClinicalTrials.gov NCT03810521.