The objective of this work is to develop a model which predicts the major factors affecting the fluorescence signal detected by a commercial fluorescence probe. The probe considered has a surface-type light-source/detector configuratiol. and utilizes the excitation and emission wavelengths corresponding to the fluorescence of NAD(P)H, which is a fluorophore found in all living material. From this model, the relationship between the fluorescence detected and the concentration of fluorescent species can be derived. Using this relationship, the probe can be used to estimate on-line the biomass concentration or the level of NAD(P)H within the cells during a fermentation. The model will also predict how other factors will influence the fluorescence signal.