Objective To investigate the potential key genes in sepsis using bioinformatics. Methods The sepsis-related gene expression data set GSE28750 was downloaded from the Gene Expression Omnibus，including 10 patients with sepsis as the experimental group and 20 healthy people as the control group. Data homogenization and differential gene screening were performed by GEO2R software （P < 0.01，fold change > 2 times）. The differentially expressed genes （DEGs） were submitted to DAVID for GO analysis and Pathway analysis，and then the identified DEGs were submitted to STRING to construct a protein-protein interaction （PPI） network to search for key genes. The survival curves of the identified key genes were plotted by GraphPad Prism 7.0 to predict the impact of key genes on the prognosis of sepsis. Results A total of 295 DEGs were screened out，including 188 up-regulated genes and 107 down-regulated genes. The GO where DEGs were enriched mainly contained immune response，small molecule metabolic process，and signal transduction. The activated signaling pathways included metabolic pathway，T cell receptor signaling pathway，and MAPK signaling pathway. The PPI network showed that the key genes CD247 and LCK were located in the core，and that the high expression groups of CD247 and LCK had higher survival rates （P < 0.05）. Conclusion The key genes CD247 and LCK are positively correlated with the survival rate of patients with sepsis. They may become new biomarkers or drug targets for sepsis，but their specific functions need to be confirmed by further studies.