Biodiesel, the mono-alkyl esters of vegetable oils or animal fats, is an eco-friendly alternative to petrodiesel. The molecular structures of biodiesels, fatty acid methyl esters were applied to predict the characteristics of biodiesel fuels. Based on the structural similarity of biodiesel and petroleum fractions, molecular weight of biodiesel was correlated with other characteristics including boiling point, viscosity and specific-gravity in the form of three equations. For 24 different kinds of biodiesel, the minimum average relative deviation (ARD) of these correlations was calculated to be 0.68%. Moreover, two correlations were developed to predict viscosity and flash point of biodiesel as a function of weighted-average number of carbon atoms (NC) and weighted-average number of double bonds (NDB) with ARD 3.72% and 4.24% respectively. Also, a high degree of correlation was shown by the logarithmic function with ARD 0.30% between specific gravity and viscosity of biodiesel. Proposed predictive models were verified by experimental data.