Background: Application of mesenchymal stem/stromal cells (MSCs) in treating different disorders, in particular osteo-articular diseases, is currently under investigation. We have already documented the safety of administrating human adipose tissue-derived stromal MSCs (hASCs) in immunodeficient mice. In the present study, we investigated whether the persistence of MSC is affected by the degree of inflammation and related to the therapeutic effect in two inflammatory models of arthritis. Methodology/Principal Findings: We used C57BL/6 or DBA/1 mice to develop collagenase-induced osteoarthritis (CIOA) or collagen-induced arthritis (CIA), respectively. Normal and diseased mice were administered 2.5×10 5 hASCs in the knee joints (IA) or 10 6 in the tail vein (IV). For CIA, clinical scores were monitored during the time course of the disease while for CIOA, OA scores were assessed by histology at euthanasia. Thirteen tissues were recovered at different time points and processed for real-time PCR and Alu sequence detection. Immunological analyses were performed at euthanasia. After IV infusion, no significant difference in the percentage of hASCs was quantified in the lungs of normal and CIA mice at day 1 while no cell was detected at day 10 taking into account the sensitivity of the assay, indicating that a high level of inflammation did not affect the persistence of cells. In CIOA mice, we reported the therapeutic efficacy of hASCs at reducing OA clinical scores at day 42 when hASCs were not detected in the joints. However, the percentage and distribution of hASCs were similar in osteoarthritic and normal mice at day 1 and 10 after implantation indicating that moderate inflammation does not alter hASC persistence in vivo. Conclusions/Significance: While inflammatory signals are required for the immunosuppressive function of MSCs, they do not enhance their capacity to survive in vivo, as evaluated in two xenogeneic inflammatory pre-clinical models of arthritis.