Production of charcoal has accompanied human life from the beginning. We aimed at evaluating the degree to which the chemical signatures of charcoal may serve as a fingerprint for burning conditions. After a compilation of fire literature we differentiated three typical fire regimes [grass and forest ground (285 ± 143 °C), shrub (503 ± 211 °C) and domestic fires (797 ± 165 °C)] and three main factors impacting on charcoal formation: charring duration, temperature and fuel. For fingerprint calibration and validation, typical fuels of prehistoric burning events (wood and grass) were charred under laboratory conditions (300–700 °C; varying duration) and compared with residues from natural fires in SE Europe. Analysis comprised assessment of benzene polycarboxylic acids (BPCAs), organic carbon (Corg) content, nitrogen content, oxygen index (OI, CO2/Corg) and hydrogen index (HI, HC/Corg), temperature of maximum heating (Tmax) and mid-infrared spectroscopy (MIRS). All parameters including mass loss increased with increasing combustion temperature, but were unaffected by charring duration. Grass charcoal had consistently lower Corg content and HI than wood, but values showed a bias towards the natural charcoals, probably because the latter contained higher amounts of mineral matter or were combusted under greater O2 supply. Nevertheless, natural charcoals could be differentiated into forest ground fires (B5CA/B6CA 1.3–1.9; OI > 20; intense CH2 stretching, Tmax < 488 °C) and grass fires (B5CA/B6CA 0.8–1.4; OI > 20; weak CH2 stretching, Tmax < 425 °C), whereas domestic fires revealed B5CA/B6CA values <0.8, OI values <20 and little MIRS absorbance. In summary, it appears possible to reconstruct fire regimes from the temperature sensitivity of BPCA patterns, Tmax, OI and aromatic and aliphatic MIRS signals, whereas assignment of fuel source was less reliable.