As the current staging system is imprecise for estimating prognosis of early stage non-small cell lung cancer (NSCLC), it is important to identify other methods for selecting high-risk patients after failed surgical treatment. The aim of the study was to evaluate the expression of 23 genes as putative prognostic markers in early stage NSCLC. The study was performed on 109 pairs of tumor and matched unaffected lung tissue surgical specimens taken from stage I and II NSCLC patients. We evaluated the mRNA level of 23 genes using the real-time PCR method. The difference in the expression between the tumor and normal tissue for each gene was analyzed using a general linear model. The influence of gene expression on survival was analyzed by using the proportional hazards model. Eighteen out of the 23 genes showed statistically significant differences in expression between the tumor and non-tumor tissue. For 12 genes (ITGB1, ITGB3, CXCL1, CXCL8, CXCL9, CXCL10, CXCL11, CXCR3, CXCR4, TNF, CHKA, AGFG1, and CTC1), the expression was lower, and for six genes (ITGA5, IL8, IL6, CXCL2, CXCL3, and CXCL12), it was higher in the tumor tissue as compared to the matched normal tissue. Expression changes were more pronounced in squamous cell carcinomas than in adenocarcinomas or large cell carcinomas. Of all the analyzed genes, only CXCL5 was found to statistically significantly (p = 0.04) influence both overall and disease-free survival. Among the 23 genes previously suggested to be relevant for early staged NSCLC patients’ postoperative outcome, only CXCL5 showed a statistically significant prognostic effect.