A 51×51 pore-throat network model was proposed to predict the primary drying stage of the freeze-drying process.The unique feature of the network model was the ability to track the evolution of drying front and phase distributions during drying.The drying curves of turkey meat and beef were simulated to verify the model.The characteristics of phase distributions during freeze drying of turkey meat samples frozen at different freezing rates were investigated.The computational characteristics of the model and the effects of pore size distributions on the phase distribution were analyzed and discussed.The simulation results showed that the model was capable of predicting the drying curves well during the primary drying stage.The pore-scale network model would provide insights into better understanding of drying dynamics at the pore level.It would serve as the theoretical foundation for rigorously determining the transition point from primary drying stage to secondary drying stage.