Objective: Currently the prevalence rate of sepsis patient with serious thrombocytopenia during the treatment of continuous renal replacement therapy (CRRT) is still high, usually considered that platelet count less than 50×109/L to continue to give invasive diagnosis and treatment, such as CRRT who need to give the corresponding treatment. This article use a retrospective study of a small samples of these patients comparing the treatments response of transfusion of fresh platelets, frozen platelet transfusion, recombinant human thrombopoietin, recombinant human thrombopoietin combined with infusion of fresh platelets, it is desirable for provided some help to clinical treatment for sepsis patients with thrombocytopenia who need uninterruptible CRRT. Methods: All 118 patients with sepsis whose peripheral platelet count decreased to ≤50×109/L after CRRT treatment in our department from January 2011 to November 2016 were selected as the research objects. According to the different treatment methods, they were divided into three groups: untreated group (23 cases), fresh platelet group (29 cases), frozen blood platelet group (25 cases), recombinant human thrombopoietin group (20 cases), and fresh platelet group (21 cases). The treatment reactions of 24 h and 72 h after treatment were compared and observed. Results: The platelet count before treatment was in different levels on the rest of the various treatment response is different. when the platelet count <30×109/L, all treatment methods were ineffective evaluated after 24 h and 72 h; the efficacy of recombinant human thrombopoietin alone and untreated group were the worst, and there were no significant differences in the other three groups curative effect; mainly therapy is platelet transfusion, lose fresh platelets and frozen apheresis platelets had no significant difference, when platelet count (30–50)×109/L, frozen platelet compared with fresh platelets and fresh platelets in the combination group treated curative efficiency was weaker after 24 h, but weaker after 72 h, and the combination group treatment was a little better than giving fresh platelets but no significant difference between the two group; using recombinant human thrombopoietin alone can slow down the platelet in a certain extent of the decline in margins; optimal treatment is transfusion of fresh platelets with or without recombinant human thrombopoietin, recombinant human thrombopoietin can also be considered with no blood products. Conclusion: Sepsis patient continued CRRT treatment with thrombocytopenia, should be based on the number of peripheral blood platelet count, to choose the different treatment option.