No matter how sophisticated the calculation method, the results can be reliable only if accurate information is supplied. The main inputs for a freezing problem, apart from the product geometry, are the heat transfer coefficient and the food’s thermal properties: freezing point, calorimetric properties, thermal conductivity and density. The freezing behaviour of foods resembles that of aqueous solutions, which start to change phase at a temperature below 0 °C (the initial freezing point) and continue to gradually freeze as temperature falls. Raoult’s law of ideal solutions is used to predict freezing point and ice fraction at various temperatures, then combined with empirical data and the bound (unfreezable) water model to develop predictive equations for the above properties at various temperatures. Recommendations are given on the choice of model for calculating thermal conductivity. Experimental properties of the Tylose gel food analogue, often used in freezing experiments, are given.