OBJECTIVE : Osteoarthritis (OA), the most encountered arthritis form, result from degeneration of articular cartilage. Obesity is accepted as a significant risk factor for knee OA (KOA). In this study, it is aimed to determine the variation of metabolites between control and patients with KOA and observe the effect of obesity on KOA via untargeted metabolomics method. METHODS : Serum samples of following groups were collected: patient group including 14 obesity (OKOA) and 14 non-obesity (NOKOA) (n = 28) and control group (n = 15) from orthopedics and traumatology policlinic. Serum proteins were denatured by acetonitrile and chromatographic separation of metabolites was achieved by LC/Q-TOF/MS/MS method. Data acquisition, classification, and identification were achieved by METLIN database. Cluster analysis was performed with MATLAB2017a-PLS Toolbox 7.2. RESULTS : Obtained results showed that 244 (patient vs control) and 274 (OKOA vs NOKOA) m/z ratios were determined in accordance with LC/Q-TOF/MS/MS analysis. Multivariate data analysis was applied 41 and 36 m/z signal (p ≤ 0.01; fold analysis > 1.5) were filtered for patient vs control group and OKOA vs NOKOA, respectively. Twenty-one different metabolites were identified for patient vs control group and 15 metabolites were determined for OKOA vs NOKOA group. CONCLUSIONS : Acid concentration and oxidative stress agents were high in inflammation group and their levels were much higher in obesity. It is claimed that obesity cause oxidative stress and acidosis in arthritis patients. Valine was found to be the only BCAA molecule whose concentration has significantly different in KOA patients. The relation between KOA and obesity was firstly investigated with metabolomics method.