Non-small cell lung cancer (NSCLC) is a common malignancy with a high morbidity and mortality rate worldwide. Herein, we employed an integrated bioinformatics approach to identify crucial genes and to investigate the underlying mechanism in the poor prognosis of NSCLC. We performed an integrated analysis of microarray datasets (GSE33532, GSE19188, GSE102287, GSE27262) containing 319 NSCLC and 232 normal lung tissues. 52 up-regulated and 173 down-regulated overlapping-differentially expressed genes were identified, further validated by cBioportal. DAVID was employed for Gene Ontology, KEGG pathway analysis, and STRING for protein-protein interaction network analysis. Cytoscape plug-ins, MCODE and CytoHubba, were used to acquire module 1–2 and to screen the top ten core genes, respectively. GEPIA and Kaplan-Meier plotter analyzed the expression levels and survival rates of the ten core genes, respectively. Four genes were up-regulated (COL1A1, MMP1, ADAM12, CTHRC1), and six genes were down-regulated (VWF, CD36, OGN, EDN1, CAV1, ITGA8) in NSCLC. Seven genes (COL1A1, ADAM12, VWF, OGN, EDN1, CAV1, ITGA8) were associated with NSCLC's poor prognosis (P < 0.01), four of which (VWF, CAV1, ITGA8, COL1A1) were enriched in the focal adhesion pathway (P = 1.04E-04), as per KEGG pathway analysis. It suggests that VWF, CAV1, ITGA8, and COL1A1 might promote NSCLC via the focal adhesion pathway.