Abstract: OBGECTIVE To evaluate the the response rate(RR),clinical benefit response(CBR) and toxicity of gemcitabine and gemcitabine plus S-1 for treating locally advanced or metastatic pancreatic cancer. METHODS Sixty-two advanced pancreatic cancer patients collected from Jan 2009-Oct 2011 in our hospital. All patients were divided into two groups, control group: gemcitabine (30 patients) and treatment group: gemcitabine plus S-1 (32 patients). In control group, patients were given gemcitabine 1000 mg·m^-2 on the 1st and 8th day, intravenous drip 30 min, repeated every 3 weeks. In treatment group, patients were given gemcitabine (usage as well as control group) and S-1 taken orally, bid on the 1-14 d, repeated every 3 weeks. RESULTS All patients were available for objective response and fifty-six(27 cases in control group and 29 cases in treatment group) patients were suitable for CBR evaluation. In control group and treatment group, RR was 23.3% and 31.3%(P<0.05), CBR was 59.2% and 72.4%(P>0.05), the 6-month survival rate was 60.0% and 68.7%(P>0.05), the 12-month survival rate was 26.6% and 31.2%(P>0.05), respective. The progression-free survival (PFS) was 3.9 months and 5.4 months(P<0.05), and the median overall survival was 7.8 months and 9.1 months(P>0.05). The incidence of toxicity between the two groups was not significant difference. CONCLUSION The two kinds of chemotherapy regimens are effective and safe for treating locally advanced or metastatic pancreatic cancer. Gemcitabine plus S-1 is better than gemcitabine in RR, it also shows superiority in survival rates, but there is no significant difference bwteen two chemotherapy regimens.