OBJECTIVE To analyze the influence of the number of lymph nodes examined on the prognosis of Dukes B and C colorectal cancer patients. METHODS The relationship between the clinicopathologic features of 373 patients with Dukes B and C colorectal cancer and number of the lymph nodes examined was retrospectively analyzed. The effect of the different number of nodes examined on the prognosis of the patients was appraised RESULTS The overall mean number of retrieved lymph nodes of the 373 patients with Dukes B and C colorectal cancer was 13.71 ±9.38. The site and size of the tumor as well as the depth of tumor infiltration were the major reasons which influenced the number of lymph nodes retrieved. The mean number of lymph nodes examined in the colon-cancer patients was 17.51 + 12.79, which was significantly more than the 11.09±6.17 (P = 0.000) examined in the rectal-cancer patients. The 5-year survival rate of the patients with Dukes B large intestinal carcinoma, with fewer lymph nodes retrieved (0 to 10), was only 60.4%, while those with more lymph node retrieved (≥10) had a 5-year survival of 77.5%. So there was a significant difference between the two groups. However the number of lymph nodes examined had no effect on prognosis of the patients with Dukes C large intestinal carcinoma. Separate analysis of the colon and rectal cancers indicated that to improve the 5-year survival rate, the number of retrieved nodes in cases with rectal cancer should be at least 9, and with colon cancer cases at least 13. CONCLUSION In order to guarantee an accuracy of tumor staging for developing a possible postoperative treatment, at least 9 lymph nodes in rectal cancer patients or 13 in colon cancer patients should be harvested.