[Objective] This study aimed to define a new subtype of lung adenocarcinoma (LUAD) and identify prognostic biomarkers; [Method] We obtained the required raw data from the TCGA and GEO database. Through gene set variation analysis (GSVA) analysis, the gene expression matrix was transformed into a gene set expression matrix. Then, the “Cancersubtypes” package in R software was used to convert the samples into different subtypes, and a LUAD-related prognosis model was established based on the differentially expressed gene sets (DEGSs) between each subtype. Finally, functional and pathway enrichment analysis, as well as a protein-protein interaction (PPI) network analysis, were performed on prognosis-related DEGSs; [Result] We obtained 63 DEGSs and constructed a prognostic model based on 4 significantly prognosis-related DEGSs; [Conclusion] This study has developed a new classification method for LUAD and identified 4 prognosis related DEGSs and 15 core protein, which could provide relevant theoretical basis and guidance for the update of cancer treatment methods and the development of new drugs.