The author's developed a practical soot model and implemented in the multidimensional computational fluid dynamics code, KIVA-3vr2 for use in low temperature diesel combustion simulations. The model framework is based on four fundamental steps: soot inception through a four-ring polycyclic aromatic hydrocarbon species, surface growth through acetylene, soot coagulation, and oxygen- and OH-induced soot oxidation. Diesel combustion was simulated by using a reduced n-heptane chemistry mechanism. A reduced polycyclic aromatic hydrocarbon chemistry mechanism was formulated from the literature and coupled with the n-heptane mechanism. Improvements were made in the chemistry mechanism for better predictions of ignition delay, liftoff length and soot precursor concentrations. The CHEMKIN-II code was used to solve the combustion chemistry. However, in order to reduce the computational time of the coupled soot and chemistry calculations, a semi-implicit solver was also implemented and used for the soot precursor species. Soot model performance and computational efficiency was evaluated by comparing the model predictions with available optical spray chamber experimental data. The model performance was also evaluated by comparing the model predictions with experimental results from a heavy-duty optical engine as well as light- and heavy-duty metal engines.