Effective separation and grading of cereal grains and crop seeds are of importance in the production of quality cereal foods. This paper presents a two-dimensional numerical study of the separation process of crop seeds by screening, using the Discrete Element Method (DEM) modelling technique. Computational experiments have been conducted for the separation of two common crop seeds, soybeans and mustard seeds, using a vibrating screen. The screening rate and the required screen length at different feeding rates are discussed in relation to the discrete particle motion on the screen. This study has demonstrated the crucial effect of particle bed depth on screening efficiency. For a screening system involving granular materials, the critical feeding rate for the most effective screening operation can be determined via conducting the DEM simulation.