Accurate and objective estimation of the postmortem interval (PMI) is crucial in forensic practice. This study aimed to infer PMI through equations based on the relationship between PMI and metabolomics biomarkers.Rats were subjected to models representing various temperatures and causes of death, with blood collected at different intervals. Untargeted gas chromatographymass spectrometry metabolomics detection methods were developed, and candidate biomarkers were chosen as co-differentially expressed metabolites in four models. A targeted method was then developed for quantitatively determining candidate biomarkers. Animal tests and human cadaver samples with clearly documented causes of death and time were used to verify the reliability of the regression equation.Results: Unique differential metabolites for CO poisoning deaths included 2,3-butanediol, hypoxanthine, and dehydrated hexanol, while those for mechanical asphyxia deaths comprised propylamine, 1,3-propylene glycol, phosphoric acid, and sorbitol. Pyruvate, glycerol and isoleucine were identified as candidate biomarkers. Human case results demonstrated the method's potential (error rate < 20%). The findings of this study may offer reference points for estimating PMI and causes of death in forensic practice.