Objective To explore the value of CD4+ T cell count and the expression rate of mHLA⁃DR in predicting the prognosis of elderly patients with sepsis. Methods One hundred and nine elderly patients with sepsis were divided into death group（n = 33）and survival group（n = 76）. The change of immunologic function indexes such as CD3+ T cell count and CD4+ T cell count and the ratio of CD4+ /CD8+ T cell count and the expression rate of monocyte HLA ⁃DR were compared and analyzed. Multivariate logistic regression was used to analyze the influencing factors affecting the 90⁃day mortality and ROC curve to analyze the potential indicators in predicting the accuracy of the 90⁃day mortality. Results CD3+ T cell count，CD4+ T cell count and the expression rate of mono⁃ cyte HLA⁃DR in death group at the 1st and the 7th day of treatment were significantly lower than those in survival group（P < 0.05）. CD4+ T cell count and the expression rate of monocyte HLA⁃DR at the 1st day of treatment were not independent risk factors for the 90⁃day mortality in sepsis. CD4+ T cell count（AUC = 0.847，95% CI：0.774 ~ 0.920）and the expression rate of monocyte HLA ⁃DR（AUC = 0.898，95% CI：0.829 ~ 0.966）at the 7th day of treatment were independent risk factors for the 90⁃day mortality and they effectively predicted the risk of death. The predicted best cut⁃off value for CD4+ T cell count was 197.0 cells/μL and that for the expression rates of monocyte HLA⁃DR 40.5%. The accuracy of the expression rate of monocyte HLA⁃DR in predicting the 90⁃day mortality was higher than that of CD4 + T cell count. The prediction specificity of the expression rate of monocyte HLA ⁃DR combined with CD4 + T cell count（AUC 0.928，95%CI 0.883 ~ 0.974）was higher than that of the single parameter. Conclusion The expression rate of monocyte HLA⁃DR and CD4+ T cell count on the 7th day of treatment areindependent risk factors and effective predictors of the 90⁃day mortality in elderly patients with sepsis and they can be targets for early intervention to improve the prognosis of sepsis.