Osteoarthritis (OA) is a significant cause of pain in both humans and horses with a high socio-economic impact. The horse is recognized as a pertinent model for human OA. In both species, regenerative therapy with allogeneic mesenchymal stem cells (MSCs) appears to be a promising treatment but, to date, no in vivo studies have attempted to compare the effects of different cell sources on the same individuals. The objective of this study is to evaluate the ability of a single blinded intra-articular injection of allogeneic bone-marrow (BM) derived MSCs and umbilical cord blood (UCB) derived MSC to limit the development of OA-associated pathological changes compared to placebo in a post-traumatic OA model applied to all four fetlock joints of eight horses. The effect of the tissue source (BM vs. UCB) is also assessed on the same individuals. Observations were carried out using clinical, radiographic, ultrasonographic, and magnetic resonance imaging methods as well as biochemical analysis of synovial fluid and postmortem microscopic and macroscopic evaluations of the joints until Week 12. A significant reduction in the progression of OA-associated changes measured with imaging techniques, especially radiography, was observed after injection of bone-marrow derived mesenchymal stem cells (BM-MSCs) compared to contralateral placebo injections. These results indicate that allogeneic BM-MSCs are a promising treatment for OA in horses and reinforce the importance of continuing research to validate these results and find innovative strategies that will optimize the therapeutic potential of these cells. However, they should be considered with caution given the low number of units per group.