A new and simple technique for the determination of total mercury in gas condensate was developed which eliminates the use of chemicals/additives and complicated digestion procedures. The determinations are carried out by vaporisation of the samples at 400 °C with adsorption of mercury species on a gold trap (Amasil) maintained at 200 °C. The trap is then heated at 900 °C to release metallic mercury, which is determined by atomic fluorescence spectrometry. The mercury recoveries from seven species, dimethylmercury (DMM), diethylmercury (DEM), diphenylmercury (DPM), methylmercury chloride (MMC), ethylmercury chloride (EMC), phenylmercury chloride (PMC) and mercury(II) chloride (MC) spiked individually into gas condensate were found to be in the range 80–100%. The mercury recoveries for mixtures of the seven species added in equal amounts to gas condensate were in the range 88–97%. For Conostan mercury standard added to the condensate, the recovery was 88%. The instrumental precision from 10 measurements of a toluene control was 4% RSD. For three mercury species. DEM, MC and EMC, added to condensate, the precision was between 2 and 5% RSD (n = 10). The limit of detection (3ς_n–1 criterion) for the procedure was calculated to be 180 pg Hg in toluene and 270 pg in condensate. For three mercury species added to a condensate sample, the absolute detection limits were 270 pg Hg for DEM, 450 pg Hg for MC and 630 pg Hg for EMC. Total mercury measurements in five real condensate samples from two sites at different stages of production covered the range 7–50 ng ml^–1 with uncertainties in the range 4–7% RSD. The total mercury concentration of two commercial heavy gas oil samples were found to be 22.2 ± 0.6 µg ml^–1 with RSD 3% (n = 4) and 2.3 ± 0.1 µg ml^–1 of mercury with RSD 3% (n = 7).