A mathematical model was developed and used to predict the performance of direct propane phosphoric acid (PPAFC) fuel cells, utilizing both Pt/C state-of the-art electrodes and older Pt black electrodes. It was found that the overpotential caused by surface processes on the platinum catalyst in the anode is much greater than the potential losses caused by either ohmic resistance or propane diffusion in gas-filled and liquid-filled pores. In one comparison, the anode overpotential (0.5 V) was larger than the cathode overpotential (0.3 V) at a current density of 0.4 A cm(-2) stop for Pt loadings 4 mg Pt cm(-2). The need for sufficient water concentration at the anode, where water is a reactant, was indicated by the large effect of H3PO4 concentration. In another comparison, the model predicted that at 0.2 A cm(-2), modern carbon supported Pt catalysts would produce 0.35 V compared to 0.1 V for unsupported Pt black catalysts that were used several decades ago, when the majority of the research on direct hydrocarbon fuel cells was performed. The propane anode and oxygen cathode catalyst layers were modeled as agglomerates of spherical catalyst particles having their interior spaces filled with liquid electrolyte and being surrounded by gas-filled pores. The Tafel equation was used to describe the electrochemical reactions. The model incorporated gas and liquid-phase diffusion equations for the reactants in the anode and cathode and ionic transport in the electrolyte. Experimental data were used for propane and oxygen diffusivities, and for their solubilities in the electrolyte. The accuracy of the predicted electrical potentials and polarization curves were normally within +/- 0.02 V of values reported in experimental investigations of temperature and electrolyte concentration. Polarization curves were predicted as a function of temperature, pressure, electrolyte concentration, and Pt loading. A performance of 0.45 V at 0.5 A cm(-2) was predicted at some conditions.