Abstract: Accurate estimation of crop yield is important for agricultural research and crop production.However, there can be a large variation of individual kernel size that can have a significant impact on final yield.Current maize models lack mechanistic processes at the individual kernel level and therefore limit their applications for a wide range of environments.Growing kernels can be separated into superior and inferior kernels.Superior kernels have the priority over inferior kernels for receiving carbohydrates.A model was developed to simulate dry matter accumulation of each kernel of an ear using a source-sink approach with a priority function.Model evaluation using experimental data from different plant densities showed that the accuracy for simulating individual kernel on different positions improved significantly when incorporating a priority function into source-sink approach.