K-RAS gene mutations have been found in 20�30% of non-small cell lung cancer and occur most commonly in adenocarcinoma, however, there was no definitive conclusion about the prognostic role of K-RAS mutations in NSCLC. Herein we performed a systematic review of the literatures with meta-analysis to assess K-RAS mutations� prognostic value in NSCLC. After a methodological assessment, survival data from published studies were aggregated. Combined hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were calculated in terms of overall survival. 41 trials (6939 patients) were included in the analysis, the overall HR was 1.45 (95% CI: 1.29�1.62), showing that K-RAS mutations have an unfavorable impact on survival of patients with NSCLC. Then a subgroup analysis was performed about ethnicity, the combined HR was 1.97 (95% CI: 1.58�2.44) for Asians, and 1.37 (95% CI: 1.25�1.5) for non-Asians. In subgroup analysis of histology, the HR was 1.39 (95% CI: 1.24�1.55) for adenocarcinoma, suggesting that K-RAS mutations were correlated with shortened survival for adenocarcinoma. When the subgroup analysis was conducted according to disease stage, K-RAS mutations were poor prognostic factors in early stages: stage I (1.81; 95% CI: 1.36�2.39) and stage I�IIIa (1.68; 95% CI: 1.11�2.55), but not in advanced stage (IIIb�IV) (1.3; 95% CI: 0.99�1.71). At last, in subgroup analysis about test methods, all of the four methods: PCR�MSOP (1.73; 95% CI: 1.35�2.2), PCR�DGGE (1.27; 95% CI: 1.01�1.62), PCR�RFLP (1.88; 95% CI: 1.42�2.49) and PCR�seq (1.34; 95% CI: 1.14�1.58) showed statistically significant impact on survival of NSCLC patients. In conclusion, this meta-analysis suggests that K-RAS mutations are associated with a worse overall survival in patients with NSCLC, especially in patients with adenocarcinoma and early stage.