OBJECTIVE: To explore the therapeutic value of pelvic lymphadenectomy for endometrial carcinoma. METHODS: One hundred and four cases undergoing pelvic lymphadenectomy were analyzed. Comparisons were carried out between various clinicopathologic factors and incidence of positive node, between 5 years survival rate with pelvic lymphadenectomy and that without pelvic lymphadenectomy. RESULTS: Four point four per cent case with positive node was found in clinical stage I cases. Positive node rate increased with deep myoinvasion and G3, accounting for 37.3% and 37.8% respectively. For clinical stage I and stage II cases, the 5 year survival rate of positive node and negative node was 38.9% and 74.2% respectively, with significant difference (P < 0.005). No significant difference was found between cases with pelvic lymphadenectomy and those without pelvic lymphadenectomy (P > 0.05) in pathologic stage I and stage II cases. CONCLUSIONS: Positive pelvic node is an important prognostic factor in endometrial carcinoma. Pelvic lymphadenectomy can not improve prognosis in early stage patients with endometrial carcinoma. It is not necessary to perform pelvic lymphadenectomy for stage I case, because of low incidence of positive node.