Background <p>Cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) pose significant global health challenges. While the mitochondrial unfolded protein response (UPR^mt) is known to influence cancer biology, its specific role in CESC remains unclear. </p> Methods <p>We employed machine learning to analyze UPR^mt genes in CESC using TCGA multi-omics data. Our comprehensive analysis included genetic alterations, prognostic significance, tumor-immune interactions, single-cell transcriptomics, pathway enrichment, and drug sensitivity assessments.</p> Results <p><i>ATF5</i> emerged as the most significant prognostic factor among UPR^mt genes, with high expression correlating with better overall survival. High <i>ATF5</i> expression was associated with an immunologically active tumor microenvironment, characterized by enhanced immune cell infiltration, increased immune checkpoint expression, and higher tumor mutational burden. Single-cell RNA sequencing revealed <i>ATF5</i>’s distinct expression patterns in stromal cells, particularly in endometrial stromal and smooth muscle cells. Gene set enrichment analysis provided mechanistic insight, revealing <i>ATF5</i>’s connection to the immune response via the regulation of P-stalk ribosome functions, a finding that underscores a novel aspect of UPR^mt’s role in shaping the tumor immune landscape. Drug sensitivity analysis showed that low <i>ATF5</i> expression correlated with resistance to conventional chemotherapeutics (cisplatin, paclitaxel, and etoposide) but increased sensitivity to imatinib, potentially through <i>EP300</i>-dependent mechanisms.</p> Conclusions <p>Our findings establish <i>ATF5</i> as both a favorable prognostic marker and a key immune response regulator in CESC. Its influence on the tumor microenvironment and treatment response suggests potential therapeutic applications. These insights into UPR^mt’s role in CESC provide new directions for developing personalized treatment strategies.</p>