Objective To explore the correlation between the scores of mental assessment techniques of image cognition and 17 items of Hamilton Depression Scale（ HAMD-17） in assessing the severity of depression. Methods A total of 113 patients with depression were retrospectively consecutively enrolled from the outpatient department of Beijing Anding Hospital. HAMD-17 was used to assess the severity of depression， and the patients were instructed to complete mental assessment based on image cognition. Partial correlation analysis and multivariate linear regression analysis were used to evaluate the correlation between image cognition score and disease severity assessed by HAMD-17 scale. The predictive value of image cognitive score on the severity of depression was analyzed by the receiver operating characteristic（ ROC） curve. Results The average score of HAMD-17 was（ 19.31±5.30）， including 18 cases of mild depression， HAMD-17 ranged from 7 to 13， with an average score of（ 11.44±1.46）； 40 cases of moderate depression， HAMD-17 ranged from 14 to 19， with an average score of（ 16.75±1.69）； 55 cases of severe disease and above， HAMD-17 over 20（ including 20）， with an average score of（ 23.78±3.03）. There were significant differences in HAMD-17 scores among patients with different degrees of depression（ F=210.99，P＜0.001）. The average score of image cognition was（ 51.02±23.22）， of which 48.7%（ 55/113） were patients with severe depression and above， and the average score of image cognition was（ 44.87±21.81）； the average score of image cognition was（ 57.06±22.91） and（ 59.35±22.82） in patients with mild and moderate depression. There was significant difference in image cognitive score among patients with different degrees of depression（ F=5.42，P＜0.01）. The image cognitive score of severe and above depression was significantly lower than that of mild and moderate depression（ all P＜0.05）. Partial correlation analysis showed that the image cognitive score was negatively correlated with the total score of HAMD-17 （r=-0.31，P ＜ 0.05）. Linear regression analysis showed that there was a linear correlation between the image cognitive score and the total score of HAMD-17（ t=-3.03，P=0.003）. The results of ROC curve analysis showed that the image cognitive score could effectively distinguish patients with major depression（ area under the curve was 0.677， P=0.001）. When the cut-off value was 52， the sensitivity was 67.2%， and the specificity was 69.1%. Conclusions The image cognitive score and HAMD-17 scale have correlation with the severity of depression， which reflects the trend of depression severity， and has a certain value in evaluating the severity of depression.