Background: COVID-19 continuously threated public health heavily. Present study aimed to investigate the lymphocyte subset alterations with disease severity, imaging manifestation, and delayed hospitalization in COVID-19 patients.Methods: Lymphocyte subsets was classified using flow cytometry with peripheral blood collected from 106 patients.Results: Multivariate logistic regression showed that chest tightness, lymphocyte count, and γ-glutamyl transpeptidase were the independent predictors for severe COVID-19. The T cell, CD4+ T cell and B cell counts in severe patients were significantly lower than that in mild patients (p = 0.004, 0.003 and 0.046, respectively). Only the T cell count was gradually decreased with the increase of infiltrated quadrants of lesions in computed tomography (CT) (p = 0.043). The T cell, CD4+ T cell, and CD8+ T cell counts were gradually decreased with the increase of infiltrated area of the maximum lesion in CT (p = 0.002, 0.003, 0.028; respectively). The T cell count, as well as CD4+ T cell, CD8+ T cell, and NK cell counts were gradually decreased with the increased delayed hospitalization (p = 0.003, 0.002, 0.013, and 0.012; respectively). The proportion of T cell was gradually decreased but B cell was gradually increased with the increased delayed hospitalization (p = 0.031 and 0.003, respectively).Conclusion: T cell and CD4+ T cell counts negatively correlated with disease severity, CT manifestation and delayed hospitalization. CD4+ T cell was mainstay of immunity response in COVID-19 patients.