Introduction <p>Mitochondrial unfolded protein response (UPR^mt) is implicated in lung adenocarcinoma (LUAD), and our study accordingly aims to establish a model incorporating UPR^mt-related genes (MRGs) for predicting the therapeutic response and prognosis in LUAD.</p> Materials and methods <p>The data were sourced from the cancer genome atlas (TCGA) and GSE31210 dataset and MRGs were retrieved to identify those with prognostic relevance, which were applied to recognize the molecular clusters in LUAD. The cluster-specific differentially expressed genes (DEGs) were identified for the functional enrichment analysis. The independent differentially expressed MRGs were sorted out to develop a risk model. Besides, the tumor immune microenvironment was analyzed using the ESTIMATE, TIMER, MCP-counter, and ssGSEA algorithms. The data were processed with Mutect2 to evaluate the genetic mutation landscape, while the IMvigor210 cohort and pRRophetic package were utilized to predict immunotherapeutic responses and drug sensitivity. Finally, in vitro validation was performed via quantitative real-time PCR (qRT-PCR), cell counting kit-8 (CCK-8), wound healing, and Transwell assays.</p> Results <p>Most MRGs were higher expressed in LUAD, and CREB binding protein (<i>CREBBP</i>), lysine demethylase 6B (<i>KDM6B</i>) and leucine rich pentatricopeptide repeat containing (<i>LRPPRC</i>) were the top 3 genes with mutation frequency. 8 MRGs were applied to identify 2 molecular clusters, with the worst prognosis seen in cluster C1. The clusters-specific DEGs were mainly enriched in cell proliferation-related pathways and the established risk model based on 4 hub genes (<i>ANLN</i>, <i>FAM83A</i>, <i>CPS1</i> and <i>KRT6A</i>) showed satisfying efficacy in predicting the prognosis and was negatively correlated with most immune cells. Besides, the tumor mutation burden tended to be stronger in high risk group with high gene mutation frequency. In IMvigor210 cohort, higher RiskScore was seen in patients with progressive disease and stable disease and related to a worse survival. 3 drug candidates, including Roscovitine, Rapamycin and PHA.665752 were positively correlated with RiskScore. Besides, all 4 MRGs were highly expressed in LUAD cells and the silencing of <i>ANLN</i> repressed the LUAD cell proliferation, migration and invasion.</p> Discussion <p>The established 4-MRGs signature not only serves as a robust prognostic indicator but also highlights the significant involvement of mitochondrial unfolded protein response in shaping tumor microenvironment and influencing immunotherapy outcomes in LUAD.</p> Conclusion <p>The 4 MRGs may contribute to the understanding on UPR^mt in LUAD and the development of relevant medicine in clinical practice.</p>