Abstract Drying is a common practice for post-harvest processing of food grains. Fluidised beds are often adopted for this purpose. It is of importance to understand the fluidised bed drying process for improving its energy efficiency. This work establishes a numerical drying model based on the combined approach of computational fluid dynamics and discrete element method for describing heat and mass transfer in the gas-solid flow system. Water evaporation is modelled in resemblance to a chemical reaction, thereby requiring fewer model parameters. The model is first described in detail. Then it is tested by comparing model predictions with those experimental data of corn kernel from the literature. General drying characteristics including grain and air moisture contents are reproduced qualitatively. The predicted drying rate curves are quantitatively comparable with those of experimental data. Finally, the effects of inlet air velocity and temperature are examined. The model predictions confirm that the drying rate increases with both the inlet air velocity and temperature. However, the drying product quality, here represented by the standard deviation of grain moisture distribution, increases with increasing air velocity or decreasing air temperature. This grain scale model would be useful to the design and control of the drying process.