Sepsis is a common and serious condition, where mitochondria and macrophage polarization play a crucial role. Therefore, this study aimed to identify and validate biomarkers for sepsis associated with mitochondria-related genes (MCRGs) and macrophage polarization-related genes (MPRGs), providing new targets and strategies for therapeutic intervention. This study utilized the GSE95233 and GSE28750 datasets. Initially, intersection genes were identified by overlapping MCRGs and the results from differential expression analysis and weighted gene co-expression network analysis (WGCNA). Biomarkers were identified through machine learning and gene expression analysis. A nomogram was developed and evaluated based on these biomarkers. Finally, functional enrichment, immune infiltration, and reverse transcription quantitative polymerase chain reaction (RT-qPCR) analyses were conducted to further elucidate the biological mechanisms underlying sepsis. The study identified YME1L1, ECHDC3, THEM4, and COQ10A as biomarkers for sepsis. Among them, YME1L1, THEM4, and COQ10A showed significantly lower expression levels in sepsis samples, while ECHDC3 exhibited markedly higher expression. Notably, RT-qPCR analysis confirmed that YME1L1, THEM4, and COQ10A exhibited significantly lower expression levels in sepsis samples. A nomogram based on these biomarkers was developed and validated, effectively predicting sepsis risk. Enrichment analysis indicated that the biomarkers were co-enriched in the oxidative phosphorylation pathway. Additionally, 13 significantly different immune cell types were identified between sepsis and control samples. Biomarker association analysis revealed that CD8 T cells had the strongest positive correlation with YME1L1 (cor = 0.84, < 0.05) and the strongest negative correlation with ECHDC3 (cor = -0.76, < 0.05), suggesting their potential role in the disease mechanism. In this study, YME1L1, ECHDC3, THEM4, and COQ10A were identified as biomarkers for sepsis, with their expression levels validated in clinical samples. These findings provided a promising theoretical foundation for the development of targeted treatments for sepsis.