Background <p>Sepsis constitutes a critical clinical condition characterized by organ failure and systemic inflammation, especially affecting children. Lipid metabolism and endoplasmic reticulum stress are involved in various diseases, yet their roles in pediatric sepsis remain unclear. This study sought to discover and validate biomarkers associated with lipid metabolism and endoplasmic reticulum stress in pediatric sepsis.</p> Methods <p>Differentially expressed genes were identified from the Gene Expression Omnibus (GEO) database. Lipid metabolism-related genes (LRGs) and endoplasmic reticulum stress-related genes (ERGs) were sourced from literature. After weighted gene co-expression network analysis and intersections, candidate genes were screened by machine learning and validated. A nomogram was built, followed by various bioinformatics analyses and quantitative reverse transcriptase PCR (qRT-PCR) confirmation.</p> Results <p>There were 12 candidate genes, and 4 of them (ACER3, DGAT2, GBA, and TSPO) were screened as biomarkers. A nomogram integrating these four biomarkers showed excellent predictive ability. Gene Set Variation Analysis (GSVA) and Gene Set Enrichment Analysis (GSEA) revealed their involvement in pediatric sepsis-related pathways. They correlated positively with stimulated dendritic cells and negatively with stimulated B cells. The molecular regulatory network included multiple RNAs and transcription factors. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) bound strongly to the biomarkers. qRT-PCR confirmed their high expression in pediatric sepsis patients’ blood (all <i>P</i> &lt; 0.05).</p> Conclusions <p>A total of 4 biomarkers (ACER3, DGAT2, GBA, and TSPO) related to lipid metabolism and endoplasmic reticulum stress in pediatric sepsis were identified. These biomarkers offered fresh insights into the mechanisms and therapeutic approaches of pediatric sepsis.</p>