Several surveys have shown temperature abuses in the last 3 steps of the cold chain: display cabinet, transport by consumer and domestic refrigerator. To evaluate the food safety, this work was carried out to predict the product temperature and microbial load evolutions in these steps. Deterministic models were used to take into account the heat transfer by convection, conduction and radiation inside the equipments. They were combined with stochastic models to take into account different sources of randomness: ambient temperature, thermostat setting, product position and residence time in the equipments etc. Distribution laws were developed to fit the survey data of these random parameters. The sampling values were used as input parameters of the heat transfer model to predict the temperature evolutions of a large number of products along the cold chain. Then, these temperature evolutions were applied to a growth model of Listeria monocytogenes to predict the bacterial contamination.