The aim of this study was to identify a panel of methylation markers that distinguish non-small cell lung cancers (NSCLCs) from normal lung tissues. We also studied the relation of the methylation profile to clinicopathological factors in NSCLC. We collected a series of 46 NSCLC samples and their corresponding control tissues and analyzed them to determine gene methylation status using the Illumina GoldenGate Methylation bead array, which screens up to 1505 CpG sites from 803 different genes. We found that 120 CpG sites, corresponding to 88 genes were hypermethylated in tumor samples and only 17 CpG sites (16 genes) were hypomethylated when compared with controls. Clustering analysis of these 104 genes discriminates almost perfectly between tumors and normal samples. Global hypermethylation was significantly associated with a worse prognosis in stage IIIA NSCLC patients (P=0.012). Moreover, hypermethylation of the CALCA and MMP-2 genes were statistically associated to a poor clinical evolution of patients, independently of TNM tumor stage (P=0.06, RR=2.64; P=0.04, RR=2.96, respectively). However, hypermethylation of RASSF1 turned out to be a protective variable (P=0.02; RR=0.53). In conclusion, our results could be useful for establishing a gene methylation pattern for the detection and prognosis of NSCLC.