Background: This study sought to investigate the association between the ERCC1/2 single-nucleotide polymorphisms (SNPs) and the efficacy of radiotherapy and prognosis in patients with non-small cell lung cancer (NSCLC). Methods: We examined 6 SNPs in the ERCC1 and ERCC2 genes in 87 consecutive patients with NSCLC who were treated with definitive radiotherapy. The objective remission rates (ORR), overall survival (OS), and progressive-free survival (PFS) were assessed. A Cox regression analysis was conducted to analyze the independent factors related to death and recurrence. Result: Patients with the G allele had better OS than patients with the A allele, and there was a statistical difference between the two groups (30.9 vs. 16.2 months; P=0.003). Patients with the AA genotype had significantly worse OS than patients with the AG or GG genotypes (6.8 vs. 19.8 vs. 30.9 months, respectively; P=0.000). The median PFS of the G allele was 18.9 months, which was significantly better than that of the A allele (P=0.040). The median PFS of patients with the GG genotype, the AG genotype, and the AA genotype was 18.9, 11.3, and 5.1 months, respectively; the difference among the three groups was statistically significant (P=0.019). Patients with the G allele also had better PFS than those with the A allele (18.9 vs. 11.3 months, P=0.040). The multivariate cox proportional hazard analysis showed that the ERCC1 gene rs11615 was an independent survival indicator [HR: 1.623, 95% confidence interval (CI): 1.018-2.591, P=0.042] but not an independent recurrence indicator (HR: 1.497, 95% CI: 0.932-2.404, P=0.095). Conclusions: The ERCC1 rs11615 SNP may be a potential biomarker for predicting survival prognosis in Chinese NSCLC patients who have undergone definitive radiotherapy. Patients with the G allele had better OS than those with the A allele.