The main objective of the research was to understand hemp processing at different stages through numerical simulations. Processing of hemp materials involves breaking the hemp into different sizes of particles and separating those particles into fractions of different sizes. Numerical models were developed using the discrete element method (DEM) to simulate hemp processing using a hammermill and separations of different hemp particles using a 3D vibratory screen-type separator. The models were implemented using a commercial DE code, the Particle Flow Code in Three Dimension (PFC3D). In the models, virtual hemp, hemp fibre and core were defined using clusters of PFC3D basic spherical particles which are connected by the PFC3D parallel bonds. The microproperties (e.g. particle stiffness and friction coefficient, and bond stiffness and strength) of these particles were calibrated. For calibrations, virtual tests were performed using PFC3D for hemp stem, fibre, and core. Those virtual tests included direct shear tests of fibre and core particles, tensile tests of fibre, and compression tests of hemp stems. The microproperties of these particles were calibrated through comparing results from the virtual tests with results from laboratory tests or literature data. Those calibrated particle microproperties were used in the PFC3D models developed for simulating the hammermill for hemp processing and the 3D vibratory separator for particle separation. These two machines were constructed using various PFC3D walls and lines, and had the main features and operational conditions as the real machines. The hammermill model was able to predict the power requirement of hammermill and particle dynamic behaviours (kinetic and strain energies) within the hammermill. The separator model was capable of predicting the separation efficiency of the 3D vibratory separator for separations of different hemp particle mixtures. The behaviour of the models reflected the real behaviour observed experimentally. The model results were reasonably good as compared with literature data and the test results. The models developed have the potential to simulate many other dynamic attributes of hemp particles with the machines. This study has laid a solid foundation for future studies of biomaterial-machine interactions using the DEM.