Objective To investigate the value of adiponectin in evaluation of the severity and CT features of chronic obstructive pulmonary disease (COPD). Methods A total of 430 patients with stable COPD who were follow-up outpatients of our department from January 2015 to December 2016 were recruited in this study. According to the severity, they were divided into 4 groups. Group A (n=116) was scored from 0 to 1 by a modified version of British Medical Research Council (mMRC) respiratory questionnaire, with the ratio of forced expiratory volume in the first second (FEV1) to the expected value ≥50%, and the frequency of acute exacerbations during last year less than twice. Group B (n=83) was defined with mMRC score ≥2, the ratio ≥50%, and the frequency less than twice. Group C (n=92) had the mMRC score from 0 to 1, the ratio <50%, or the frequency more than twice. Group D (n=139) was assigned as mMRC score ≥2, the ratio <50%, or the frequency more than twice. Another 206 healthy subjects with normal pulmonary function were enrolled as control group. All underwent pulmonary function test, and high-resolution computed tomography (HRCT) for the percentage of lung voxels with low-attenuation areas (LAA%), the ratio of 2-fold airway wall thickness to outer diameter (2T/D) and the ratio of wall area to total airway area (WA) were mainly tested. Plasma adiponectin level was measured by enzyme-linked immunosorbent assay (ELISA). Results There were significant differences in fasting blood glucose (FBG), hemoglobin A1c (HbA1c), adiponectin, FEV1/FVC, LAA%, 2T/D, WA%, FEV1/Expected value (%), diffusing capacity of the lungs forcarbon monoxide (DLCO), and ratio of residual volume (RV) to total lung capacity (RV/TLC) among the 4 groups (P<0.05). The results of univariate analysis showed that gender, body mass index (BMI), the frequency of acute exacerbations in last year, lung function and LAA classification significantly affected the level of adiponectin (P<0.05). Correlation analysis indicated that plasma adiponectin levels in COPD patients were positively correlated with RV/TLC (r=0.002), LAA% (r=0.010), 2T/D (r=0.006), WA (r=0.011), smoking index (r=0.356), the score of COPD assessment test (CAT) (r=0.497), the frequency of acute exacerbations in last year (r=0.749) (P<0.05), and negatively correlated with BMI (r=-0.440), DLCO (r=-0.528), FEV1/FVC (r=-0.247), and FEV1/Expected value (r=-0.037, P<0.05). Conclusion Adiponectin may be used as a phenotype marker for pathological changes of COPD with different severity.