A fragment-based methodology for estimating the thermophysical properties of triglycerides is presented. In contrast to the commonly practiced functional group estimation approach, the proposed methodology adopts a chemical constituent fragment-based approach to estimate the triglyceride pure component properties from fragment composition and parameters of the fragments. The fragment-specific parameters are obtained from regressing against very limited experimental data for triglycerides available in the literature. The methodology further explores the relationships between carbon atom numbers of fatty acid constituent fragments and the values of these fragment-specific parameters. Additionally, the effect of the double bonds on the values of these fragment-specific parameters is investigated. Based on this methodology, we develop the first-ever pure component thermophysical property databank of triglycerides. We show satisfactory predictions on the properties of triglycerides, fats, and oils. We further show the superiority of this methodology over the traditional functional group approach. The methodology, the derived databank, together with the currently available databanks for fatty acids and corresponding esters, enable efficient and reliable thermophysical property calculations in support of process modeling, simulation, design, and optimization of biodiesel production processes.