Obstructive sleep apnea (OSA) is prevalent in obese subjects. Plasma adiponectin level in obese subjects is decreased. Whether reduced adiponectin level is associated with OSA is unknown. Participants without a previous diagnosis of OSA or who have not been treated with continuous positive airway pressure were enrolled and parameters of interest were collected. Polysomnography was performed to evaluate the presence of OSA and the severity of OSA as indexed by the apnea-hypopnea index (AHI). Between-group differences were analyzed. Pearson correlation analysis was used to evaluate the association between body mass index (BMI) with plasma levels of adiponectin and C-reactive protein (CRP) and AHI; and the association between plasma adiponectin level with CRP and AHI was also evaluated. Logistic regression analysis was conducted to evaluate the association between per 1-SD standardized decrease of plasma adiponectin level and the prevalence of OSA using stepwise adjustment models. A total of 486 participants were enrolled and the mean BMI was 26.9 ± 6.2 kg/m2 with obesity prevalence of 28%; and the mean AHI was 12.6 ± 8.9 per sleep hour with OSA prevalence of 42%. The mean adiponectin level was 18.4 ± 10.6 μg/mL. Compared with the nonobese group, participants in the obese group had higher BMI, neck girth, waist circumference, and AHI (P < .05 for all comparisons). The prevalence of OSA (51% vs 37%) and the proportion of moderate OSA (49% vs 42%) were also significantly higher, while adiponectin level (14.6 ± 8.7 μg/mL vs 20.7 ± 10.5 μg/mL) was significantly lower. In the obese group, plasma adiponectin level was decreased gradually with the increasing severity of OSA, which was not observed in the nonobese group. BMI was negatively correlated with adiponectin while positively correlated with CRP and AHI; and adiponectin was negatively correlated with both CRP and AHI. After adjusted for covariates including BMI and waist circumference, adiponectin remained significantly associated with OSA prevalence with odds ratio of 1.20 (95% confidence interval 1.12–1.65). In summary, our preliminary study suggests that in obese subjects, plasma adiponectin level is associated with the prevalence of OSA.