Driven by the need to model formation of detailed species from hydrocarbon fuel combustion in multidimensional simulations, this study proposes a skeletal kinetic mechanism to predict experimentally measured aromatics and soot precursors formed in nonpremixed flames of propane. As existing kinetic mechanisms of propane oxidation are either oversimplified or too complex to be used for any computational fluid dynamics (CFD) model, the computational procedure begins with combining a 33-species mechanism of propane and a 145-species submodel of polycyclic aromatic hydrocarbons (PAHs) to broaden the representation of chemical kinetics. The mechanism reduction is performed to generate a minimized but functionally equivalent mechanism consisting of 75 species and 327 reactions. Without empirical adjustment of kinetic parameters in elementary reactions, the newly derived propane-PAH mechanism is refined for the prediction of nonfuel products, including 11 straight-chain unsaturated hydrocarbons and 13 cyclic compounds. Incorporated into a 2-D axisymmetric laminar finite-rate model, the propane-PAH mechanism reproduces mole fraction peaks and centerline profiles in accordance with measured data in a nonpremixed flame of propane, which are computationally investigated for the first time. Furthermore, the concentration contours and rate-of-production analysis reveal the reactions correlated with the fuel decomposition and formation of nonfuel products. In all, the newly proposed C3H8-PAH mechanism reduces the gap between fuel chemistry and combustion physics relating to CFD simulations.