Computer models are a rapid, inexpensive way to identify agricultural areas with a high potential for P loss, but most models poorly simulate dissolved P release from surface-applied manures to runoff. We developed a simple approach to predict dissolved P release from manures based on observed trends in laboratory extraction of P in dairy, poultry, and swine manures with water over different water to manure ratios. The approach predicted well dissolved inorganic (R2 = 0.70) and organic (R2 = 0.73) P release from manures and composts for data from leaching experiments with simulated rainfall. However, it predicted poorly (R2 = 0.18) dissolved inorganic P concentrations in runoff from soil boxes where dairy, poultry, and swine manures had been surface-applied and subjected to simulated rainfall. Multiplying predicted runoff P concentrations by the ratio of runoff to rainfall improved the relationship between measured and predicted runoff P concentrations, but runoff P was still overpredicted for dairy and swine manures. We attributed this overprediction to immediate infiltration of dissolved P in the freely draining water of dairy and swine manure slurries upon their application to soils. Further multiplying predicted runoff dissolved inorganic P concentrations by 0.35 for dairy and 0.60 for swine manures resulted in an accurate prediction of dissolved P in runoff (R2 = 0.71). The ability of our relatively simple approach to predict dissolved inorganic P concentrations in runoff from surface-applied manures indicates its potential to improve water quality models, but field testing of the approach is necessary first.