Abstract. The present study aimed to evaluate the effectiveness of erythropoietin (EPO) for improving cancer-associated malignant anemia. A search was performed for randomized clinical trials, conducted according to the Cochrane manual, using electronic databases including PubMed, EMBASE, the Cochrane Library and ClinicalTrails.gov up to 15 August 2015. A total of 6 eligible studies from 5 articles enrolling a total of 453 patients were entered into the current meta-analysis. Upon EPO treatment, there were significant differences in the change in hemoglobin (HB) levels compared with the placebo at short-term follow-up [mean difference (MD)=0.66; 95% confidence interval (CI), 0.14-1.18; I 2 =Not applicable; P=0.01) and long-term follow-up (MD=0.10; 95% CI, 0.02-0.18; I 2 =Not applicable; P=0.01) under the random effects model. For changes in hematocrit (HCT) compared with the placebo, the results revealed there were significant differences at short-term follow-up (MD=2.47; 95% CI, 0.75-4.19; I 2 =Not applicable; P=0.005) and long-term follow-up (MD=7.60; 95% CI, 6.15-9.05; I 2 =Not applicable; P<0.00001) under the random effects model. Compared with the placebo in short-term follow-up under the fixed effects model with homogeneity, the result was a significant difference for the transfusion ratio [relative risk (RR)=0.81; 95% CI, 0.67-0.97; I 2 =34%; P=0.02) and the transfusion requirements (MD=-0.45; 95% CI, -0.92, 0.03; I 2 =6%; P=0.07). Funnel plots did not detect any publication bias. These results suggest that EPO was beneficial to alleviate cancer-associated anemia and improve survival outcomes for patients with cancer.