Despite the high incidence and disability rates of delayed encephalopathy after acute carbon monoxide poisoning (DEACMP), its pathogenesis remains enigmatic, and specific predictive markers are lacking. This study aimed to elucidate the molecular underpinnings and identify predictive biomarkers of DEACMP through multi-omics and single-nucleus RNA sequencing (snRNA-seq). We collected clinical data and blood samples from 105 participants, including healthy controls (HCs), acute carbon monoxide poisoning patients (ACOP), and those receiving comprehensive treatment for ACOP (ACOP-CT). Untargeted metabolomics sequencing was employed to profile serum metabolites across these groups. Additionally, individuals from the HCs, ACOP, non-delayed encephalopathy after ACOP (DEACMP-N), and DEACMP groups (n = 3 each) were randomly selected for transcriptome sequencing to identify potential predictive targets and pivotal signaling pathways associated with DEACMP. Furthermore, we established severe DEACMP and Control Sprague-Dawley rat models and assessed neurocognitive function using the Morris water maze on the 28th day. Subsequently, three rats from the Control, DEACMP, and DEACMP + Dexamethasone + Selenomethionine groups were selected for snRNA-seq to analyze hippocampal single-cell transcriptional profiles. Immunofluorescence multiplexing was then performed to validate the identified predictive targets. Our analysis of clinical data from 105 participants highlights the pivotal role of inflammation in influencing the prognosis of carbon monoxide poisoning. Metabolomics analysis identified 19 metabolites that significantly differed between the DEACMP-N and DEACMP groups compared to the ACOP-CT follow-up results. Transcriptomics analysis of 12 participants indicated that DEACMP is primarily associated with six signaling pathways, including lysosome and tuberculosis. Given that microglia are central nervous system immune effectors, our snRNA-seq analysis revealed altered genes expression and signaling pathways in microglia during DEACMP, with KEGG analysis highlighting phagosome, neutrophil extracellular trap formation, lysosome, and tuberculosis as the predominant pathways. Differential gene analysis from transcriptome and snRNA-seq identified 28 genes differentially expressed in DEACMP. The STRING database and immunomultiplexing confirmed the pivotal role of the IFNGR1/STAT1/CTSS axis in DEACMP. This study provides a comprehensive overview of serum metabolite expression, differential genes expression, and signaling pathways in DEACMP patients, offering a robust theoretical foundation for understanding the pathogenesis for DEACMP.