© 2019 American Chemical Society. We develop four kinetic models of varying complexity for biodiesel production. The models incorporate both transesterification and saponification, thereby making them practically applicable. We then propose an iterative parameter estimation algorithm to identify a prefixed number of significant rate constants via sensitivity analysis and estimate their kinetic parameters (A and ΔE) using nonlinear regression. Using experimental data on eight different oils, two alcohols, and two catalysts, we show that our models accurately predict the dynamic concentration profiles of various species during the transesterification of oil. Furthermore, we demonstrate the applicability of the best model (based on the values of Mean Absolute Error, Root Mean Square Error, and Akaike Information Criterion) for 11 additional experiments by predicting the final biodiesel properties with significant accuracy. Finally, using N-way ANOVA, we identify the choice of oil, alcohol, and catalyst as the most significant input factors followed by the operating conditions of the reactor.