BACKGROUND AND OBJECTIVE The incidence of intrahepatic cholangiocarcinoma (ICC) is low. Current treatment for ICC is unsatisfied. This study was to investigate the prognosis of patients with resectable or unresectable ICC. METHODS Clinical data of 84 patients with pathologically confirmed ICC treated at Cancer Center, Sun Yat-sen University from January 1997 to December 2007 were reviewed. Survival and prognosis were analyzed by Kaplan-Meier method and Cox regression model. RESULTS Of the 84 patients, 56 (66.7%) had resectable ICC, and 28 (33.3%) had unresectable ICC. Among the 56 patients with resectable ICC, 27 (48.2%) underwent radical resection, and 29 (51.8%) underwent palliative resection. The 2-year overall survival rate was 3.1% in unresectable ICC group; it was significantly higher in radical resection group than in palliative resection group (P<0.01). For the patients with resectable ICC, univariate analysis revealed that operation pattern, histological type, tumor size and number, lymph node metastasis, intrahepatic metastasis, portal vein thrombus, postoperative serum level of albumin, preoperative serum levels of CEA, CA199, TBIL, ALT and AST were related to the prognosis; multivariate analysis found that operation pattern, histological type, tumor number, preoperative serum levels of CEA, CA199 and TBIL were independent prognostic factors. For the patients with unresectable ICC, univariate analysis found that histological type and preoperative serum level of CA199 were related to the prognosis; whereas multivariate analysis found that histological type was the only independent prognostic factor. Chemotherapy showed no survival benefit in both resectable and unresectable ICC groups (P=0.30, P=0.78). CONCLUSIONS Radical resection is the main effective treatment for ICC patients to achieve long-term survival. Preoperative serum levels of CEA, CA199 and TBIL are significant prognostic factor for patients with resectable ICC.