Multivariable models were developed to predict the melting points of alkanes having 10-20 carbon atoms and only one methyl group. These types of alkanes are of special interest to petroleum engineers manufacturing synthetic diesel fuel. Models were based on the number of carbon atoms, the Wiener Path numbers, the mean Wiener index, and the methyl locant index. Nonlinear regression techniques were used which employed web-based NEOS solvers to determine the coefficients and exponents of the indices in the models. The models were used to predict the melting points of alkanes having 10-20 carbon atoms for which no experimental melting-point data existed.