To develop a GC-MS method for the detection of paraquat by sodium borohydride-platinum chloride reduction in postmortem samples. Paraquat [1,1′-dimethyl-4,4′-bipyridilium dichloride,(PQ)] is a bipyridylium quaternary ammonium contact herbicide. At the Institute of Forensic Science and Legal Medicine (IFSLM), an average of 13 cases of PQ poisoning are submitted annually for postmortem toxicological analysis. Previously a colorimetric method for the identification of PQ was used. A more sensitive method was needed to confirm the presence of PQ. There is limited studies on liquid-liquid extraction (LLE) of postmortem fluids and tissues for determination of PQ by GC-MS. Some methods used a chemical reduction involving sodium borohydride (NaBH4) or sodium borohydride with nickel chloride (NaBH4-NiCl2), prior to LLE, to gain more volatile compounds for analysis. In the absence of NiCl2, platinum (IV) chloride (PtCl4) was used as a substitute catalyst in the validation. The aim of this validation is to identify PQ in suspected poisoning cases from postmortem samples using chemical reduction reaction, followed by LLE and then (GC-MS). The qualitative method involved an initial chemical reduction of PQ to the perhydrogenated product and monoene by-product: 500 μL blood, urine or tissue homogenate was placed in a glass tube, followed by 200 μL toluene, 400 μL PtCl4 (1% w/v) and then 500 μL NaBH4 (40% w/v) was added dropwise. The mixture was vortexed then left to stand for 1½ hours at room temperature. 1 mL 5 N sodium hydroxide solution was added to stop the reaction. For LLE, 3 mL hexane: ethyl acetate (4:1) was added to the tube and placed on shaker for 45 minutes at 300 rpm. The sample was centrifuged for 1 hour at 1400 g. The hexane:ethyl acetate layer was then transferred to a beaker and allowed to evaporate to dryness. The residue was reconstituted with 1 mL hexane, transferred to GC vial, and allowed to evaporate to dryness. The residue was then reconstituted with 100 μL hexane, and transferred to a GC vial fitted with glass insert, with 1 μL being injected into the GC-MS system (Agilent 7890A GC coupled with an Agilent 5975C VL MSD with Triple Axis Detector Mass Spectrometer). As recommended by the Scientific Working Group for Forensic Toxicology (SWGTOX), the validation parameters for qualitative confirmation/identification were specificity, recovery, carryover, repeatability and limit of detection. The method was applied to blank blood and urine samples, spiked blood and urine samples, as well as 10 suspected PQ poisoning cases (24 postmortem samples) submitted to the IFSLM. The blank blood and urine samples analysed did not yield any positive results for the analytes of interest. Blood and urine samples spiked with the PQ standard solution resulted in the detection of both the perhydrogenated product and monoene by-product. The limit of detection for both reduction products was 30 ppm. There was no detection of the analytes in the three blanks analysed after an injection of 1000 ppm PQ. The reduction products of PQ were detected in all 10 cases. Four of these cases had previously given negative results using the colorimetric test. The PtCl4 was found to be an adequate substitute catalyst in the reduction procedure. The procedure was proven to be sensitive, repeatable and robust. All sample types assessed by the method gave positive results, however stomach and content produced the best chromatograms with sharp, clearly defined and resolved peaks. In summary, the presented procedure provided excellent performance in the identification of the reduction products of PQ in postmortem samples.