Eggs have become main sources of protein choice for Chinese consumers due to the fact that they are both inexpensive and rich in vitamins, minerals and protein. However, as a perishable product, the quality of fresh eggs deteriorates continuously during the period from their leaving the farm until final consumption or use in manufacturing. With consumers’ increasing awareness and concern for food safety, increasing attention is being given to the shelf life of eggs through the supply chain. To develop a prediction model of the shelf life of fresh eggs, two types of model were developed and tested, including a kinetic model and a back-propagation (BP) neural network model. A sample of 115 eggs was collected on the same day from the same farm layer-hen house subsequently for use in simulating quality deterioration under laboratory conditions. The experiments were conducted at constant temperatures of 5, 25 and 35℃ to cover the normal range of temperatures that can occur under real egg storage conditions and the experimental results were used to construct the kinetic and BP neural network models, and validation of model shelf-life prediction was compared with actual egg shelf life. Three layers of BP neural network were constructed with Haugh units, yolk index and temperature as the input layer parameters, 10 nodes in the hidden layer and remaining day’s duration of storage as the output layer’s parameter. It was found that the BP neural network model had a superior prediction accuracy of 95.93% compared with 90.79% of the kinetic model. Hence it can be concluded that the BP neural network model could readily be integrated as part of a quality control system setting sell or use-by-dates for consumers.