OBJECTIVE： To investigate the correlation between ribonucleotide reductase M1 subunit （RRM1） single nucleotide polymorphisms （SNPs） and chemotherapy sensitivity of patients with non-small cell lung cancer （NSCLC） for gemcitabine. METHODS： A total of 96 NSCLC patients receiving primary treatment selected from our hospital during Aug. 2014-Jul. 2016 were all accepted gemcitabine-based two-drug chemotherapy plan， with continuous treatment for at least 2 cycles （28 d as a cycle）. Chemotherapy sensitivity rate was calculated by using the ratio of the sum of patients with complete response and partial response to the sum of test patients. RRM1 genotype was tested by PCR and direct sequencing. The correlation between different genotypes and chemotherapy sensitivity was analyzed. RESULTS： Distribution frequency of RRM1-37C＞A CC， CA， AA genotype were 35.42％， 52.08％， 12.50％， respectively； distribution frequency of -524C＞T CC， CT， TT genotype were 18.75％， 37.50％， 43.75％， respectively. The frequency of each genotype was in the line with Hardy-Weinberg equilibrium （P＞0.05）. Chemotherapy sensitivity rate of 96 NSCLC patients was 37.50％. The patient’s age， sex， ethnicity， smoking or not， TNM stage， pathological type， chemotherapy plan， and the Eastern American Oncology Collaboration score were not associated with chemotherapy sensitivity （P＞0.05）. Chemotherapy sensitivity rates of RRM1（-37CA）+（-524CT）genotype and （-37CC）+（-524TT） genotype patients （57.14％， 39.39％） were significantly higher than those of other genotype patients （10.71％）， with statistical significance （P＜0.05）. There was no statistical significance in chemotherapy sensitivity rate between RRM1（-37CA）+（-524CT） and （-37CC）+（-524TT）genotype patients. CONCLUSIONS： In NSCLC patients， the SNPs of RRM1 can be used as predictive factor for the sensitivity of gemcitabine chemotherapy， and RRM1（-37CA）+（-524CT） and（-37CC）+（-524TT） genotype patients have higher sensitivity to this type of chemotherapy.