Objective The pathophysiology of social anxiety disorder (SAD) is not yet well understood, but previous research has suggested that oxytocin is associated with social behavior and may play a role in human anxiety states and anxiety-related traits. The aim of this study was to investigate the possible relationship between social anxiety symptoms and plasma oxytocin levels. Methods Twenty-three male patients with SAD and 28 healthy male controls participated in this study. All participants were assessed using the Mini International Neuropsychiatric Interview (MINI) and the Liebowitz Social Anxiety Scale (LSAS). Multivariate regression analysis was performed to identify associations between plasma oxytocin levels and SAD. Results In multiple regression models, after controlling for age and years of education, we found that higher oxytocin levels were significantly associated with higher total LSAS scores (R2=0.157, coefficient=0.145, 95% CI=-0.0005–0.291, p = 0.051) and fear subscale scores (R2=0.134, coefficient=0.083, 95% CI=0.007–0.159, p=0.034) in the SAD group. Conclusion In this study, increased plasma oxytocin levels were associated with higher social anxiety symptoms among SAD patients, but not among controls. This might be because among SAD patients, higher oxytocin (OT) secretion is an insufficient compensatory attempt to reduce social anxiety symptoms.