Pulsed light (PL) technology is able to effectively destroy a wide variety of food spoilage and pathogenic microorganisms. However, the effectiveness of PL treatment depends on direct exposure of the target microorganisms to the short, high energy pulses of light. The complex physical and chemical properties of foods affect the way light propagates through a given food substrate, and thus there is a real potential for insufficiently or non-uniformly treated products. The objective of this work was to develop a method for predicting levels and spatial distribution of microbial inactivation in PL treatment of liquid substrates, and to validate the predictions with experimental data. Three liquids with different composition and optical properties (BPB, TSB, apple juice) were inoculated with either Escherichia coli ATCC 25922 or Listeria innocua FSL C2-008 and treated with PL, in two different geometries. The Weibull model was used to describe the microbial inactivation kinetics for each organism. The kinetic equations were coupled with previously determined equations describing either the total fluence (F(total)) or UV fluence (F(UV)) distribution in each of the liquids, for either cylindrical or rectangular prismatic geometries. COMSOL simulation software was used to generate maps of spatial distribution of microbial inactivation and to predict the average volumetric inactivation for each substrate. The model that used F(total) provided gross over-estimations for microbial inactivation, while using F(UV) as the treatment dose yielded reasonably good predictions of microbial inactivation, especially for the more opaque and turbid substrates. This approach can help processors determine which substrates would be suitable for PL treatment, and to design highly effective and uniform PL treatments. (C) 2011 Elsevier Ltd. All rights reserved.