A new procedure was developed for predicting the correction factor of a deviant thermal process applied to canned or packaged convection heating food. This development was accomplished through a combined application of a mathematical model for estimating process lethality and statistical procedures. Through a dimensional analysis of the mathematical model it was found that there were nine independent dimensionless parameters required to define uniquely correction factors. The influence of each parameter was determined by applying a statistical design of screening tests. Of the nine dimensionless parameters, only two parameters, which contained a magnitude of drop in heating medium temperature, duration of this drop, the slope index of thermal death time curve, and the slope index of heating curve influenced significantly the correction factor. Those of no significant influence included initial food temperature, retort temperature, cooling water temperature, drop location, and normal processing time. A simple polynomial equation was obtained for predicting the correction factor by applying a central composite design of experiments and a statistical regression analysis. The reliability of this equation was verified through heat transfer experiments by using canned distilled water, 0.1% bean gum aqueous solution, and tomato juice.