Most cosmetic products are susceptible to microbiological spoilage due to contaminations that could happen during fabrication or by consumer's repetitive manipulation. The composition of cosmetic products must guarantee efficient bacterial inactivation all along with the product shelf life, which is usually assessed by challenge-tests. A challenge-test consists in inoculating specific bacteria, i.e. Staphylococcus aureus, in the formula and then investigating the bacterial log reduction over time. The main limitation of this method is relative to the time-consuming protocol, where 30 days are needed to obtain results. In this study, we have proposed a rapid alternative method coupling High Content Screening-Confocal Laser Scanning Microscopy (HCS-CLSM), image analysis and modeling. It consists in acquiring real-time S. aureus inactivation kinetics on short-time periods (typically 4h) and in predicting the efficiency of preservatives on longer scale periods (up to 7 days). The action of two preservatives, chlorphenesin and benzyl alcohol, was evaluated against S. aureus at several concentrations in a cosmetic matrix. From these datasets, we compared two secondary models to determine the logarithm reduction time (Dc) for each preservative concentration. Afterwards, we used two primary inactivation models to predict log reductions for up to 7 days and we compared them to observed log reductions. The IQ model better fits datasets and the Q value gives information about the matrix level of interference.