Temperature controlled food transport is essential for human safety and to minimise food waste. The thermal properties of food are important for determining the heat transfer during the transient stages of transportation (door opening during loading and unloading processes). For example, the temperature of most dairy products must be confined to a very narrow range (3–7 °C). If a predefined critical temperature is exceeded, the food is defined as spoiled and unfit for human consumption. An improved empirical model for the thermal conductivity and specific heat capacity of a wide range of food products was derived based on the food composition (moisture, fat, protein, carbohydrate and ash). The models that developed using linear regression analysis were compared with the published measured parameters in addition to previously published theoretical and empirical models. It was found that the maximum variation in the predicated thermal properties leads to less than 0.3 °C temperature change. The correlation coefficient for these models was 0.96. The t-Stat test (P-value >0.99) demonstrated that the model results are an improvement on previous works. The transient heat transfer based on the food composition and the temperature boundary conditions was found for a Camembert cheese (short cylindrical shape) using a multiple dimension finite difference method code. The result was verified using the heat transfer today (HTT) educational software which is based on finite volume method. The core temperature rises from the initial temperature (2.7 °C) to the maximum safe temperature in ambient air (20.24 °C) was predicted to within about 35.4 ± 0.5 min. The simulation results agree very well (+0.2 °C) with the measured temperature data. This improved model impacts on temperature estimation during loading and unloading the trucks and provides a clear direction for temperature control in all refrigerated transport applications.