A mathematical model has been developed for the prediction of sample temperature, average moisture and moisture distribution in a cylinder-shaped solid during the drying process. The effect of shrinkage was taken into account. The macroscopic heat balance and the microscopic mass balance combined with Fick＇s law were simultaneously solved using the Runge-Kutta-Merson method and a numerical finite difference method. The effective diffusion coefficient was expressed as a function of sample temperature and local moisture content. Using an experimental drying curve determined at 90 degrees C, the diffusional equation was identified for broccoli stems, and was used to predict the average and local moisture content variations during the process. Good agreement with experimental results was found (average percentage of explained variance was 99.8%).