Adipogenesis and osteogenesis, a reciprocal relationship in bone marrow, are complex processes including proliferation of precursor cells, commitment to the specific lineage, and terminal differentiation. Accumulating evidence from in vitro and in vivo studies suggests that melatonin affects terminal differentiation of osteoblasts and adipocytes, but little is known about the effect of melatonin on the process of adipogenesis and osteogenesis, especially adipogenesis. This study was performed to determine the effect of melatonin on adipogenesis and osteogenesis in human mesenchymal stem cells (hMSCs). Cell proliferation assays demonstrated that melatonin had no apparent effect on the proliferation of hMSCs. When melatonin was added to the adipogenic/osteogenic medium, it directly inhibited adipogenesis and simultaneously promoted osteogenesis of hMSCs in a dose-dependent manner. Furthermore, quantitative RT-PCR demonstrated that melatonin significantly suppressed peroxisome proliferator-activated receptor gamma (PPARγ) expression (day 3, 25% decrease; day 6, 47% decrease), but promoted Runx2 expression (day 3, 87% increase; day 6, 56% increase) in the early stages of adipogenesis and osteogenesis of hMSCs. Moreover, melatonin down-regulated several markers of terminal adipocyte differentiation, including leptin (30%), lipoprotein lipase (LPL, 41%), adiponectin (51%), and adipocyte protein 2 (αP2, 45%). Meanwhile, melatonin up-regulated several markers of osteoblast differentiation, including alkaline phosphatase (110%), osteopontin (218%), and osteocalcin (310%). These results suggest that melatonin directly inhibits hMSCs adipogenic differentiation and significantly enhances hMSCs osteogenic differentiation by suppressing PPARγ expression and enhancing Runx2 expression; this provides further evidence for melatonin as an anti-osteoporosis drug.