This paper presents the application of artificial neural networks to adiabatic flame temperature prediction of hydrocarbon fuels. The investigation was conducted over a wide range of operating conditions in terms of fuel composition, pressure and temperature of reactants, fuel-air equivalence ratio and fuel vapour fraction. Several neural network models for predicting the flame temperature for different applicable fuel ranges were built and examined. The proper preparation of network training data and the appropriate choice of network parameters for achieving better prediction accuracy are discussed. The neural network prediction results were compared with those calculated by a thermodynamic and chemical equilibrium-based computer code-the NASA program CET89. It was shown that trained neural network models can provide the adiabatic flame temperature prediction with a good level of accuracy over a wide range of operating conditions.