Explore the possibility of using postmortem metabolic fingerprinting for determine cause of death. We have previously demonstrated that the low-weight molecules present in a biological system after death has the potential to reflect the events leading up to death. That molecular overview (i.e. the postmortem metabolome) could potentially be used as a fingerprint useful in determining the cause of death. In Sweden, approximately five thousand forensic autopsies are carried out each year where approximately fifteen hundred cases have died either from drug intoxication ( n ≈800), strangulations by hanging ( n ≈500), pneumonia ( n ≈150) or acidosis ( n ≈50). The aim of this study was to investigate if our toxicological screening could be used to generate a postmortem metabolic fingerprint of autopsy cases that died either from drug intoxication, strangulation by hanging, pneumonia or acidosis. All autopsy cases handled at the Swedish National Board of Forensic Medicine between June 2017 and October 2020, in which the cause of death was opioid and benzodiazepine intoxications, hanging, pneumonia or acidosis, with a toxicological screening in femoral blood, with a BMI between 18,5–30 kg/m 2 , with an age between 30–65 years, were considered for this study. From each of the four causes of death, twenty-nine autopsy cases were randomly selected ( n = 116). The ultra-high performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UHPLC-QTOF) data were processed using XCMS (R) and evaluated with principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA). Three quarters of the samples ( n = 88) were used for model building while the remaining quarter ( n = 28) were used for external model validation. A simplified model was built with the most pronounced and significant chromatographic peaks. The chromatographic peaks were identified using the online databases METLIN and HMDB. For the training set, with 1355 chromatographic peaks as x-variables a full group separation was observed and with a 100% correct classification of intoxication, hanging, pneumonia and acidosis cases. The model had a R2 and Q2 value of 0.88 and 0.49 respectively, indicating that the model described the data well and that the model was reproducible. For the validation set, 82% of the samples were correctly classified. In a simplified model, only based on 20 chromatographic peaks (e.g. acylcarnitines, aminoacids, ketonbodies, polyethyleneglycols, steroid hormones), 72% of the cases in the validation set were correctly classified. This study demonstates that postmortem metabolic fingerprinting could potentially be used as a quick screening approach in search of plausible cause of death which could be a vital addition in death investigations. Further, several of the identified chromatographic peaks seemed to be biologically relevant. For example, increased levels of cortisone and decreased levels of LysoPC observed in the pneumonia cases could reflect the severity of illness and the prolonged agonal period expected in pneumonia cases in comparison to the other cases. Furthermore, decreased levels of acylcarnitines were observed in the intoxication cases which could reflect the respiratory function and the mitochondrial activity, being an indicator for respiratory depression. Postmortem metabolic fingerprints is a promising tool in death investigations, but more studies are needed to fully explore its potential.