Granular grinding is one of the most important unit operations used in a wide variety of industries. Examples can be found in the food industry, for instance, rice processing, etc.. The performance of grinding can be characterized by the particle flow process. Thus in order to study the stable flow process of particles during grinding, we must establish a discrete element model (DEM) of granular axial flow in the grinding area between the grinding roller and the screen drum. DEM is a numerical method used for modelling the mechanical behaviour of granular materials. When DEM is used in grinding, the particle motion is controlled by contact models that are governed by physical laws. Using EDEM software, the process of grinding can be simulated and analyzed. The simulation system chooses continuous feeding; after a period of time, it reaches a steady flow. Research results show that the uneven distribution of particle flow density (PFD) is caused by the axial movement difference of particles in the grinding area. The form, flow rate and distribution of granular axial flow are influenced by static friction coefficient difference between particles and screen drum. Axial mean square deviation of single particles in the grinding area is positively correlated with the square of time, which follows a “super” diffusive behavior defined by some studies. By an overall consideration of the grinding area, we find that the axial average velocities increase, however, the average velocities that are synthesized by three-axis velocities gradually decrease along the axial direction. This is because in a different axial position with different PFI, the PFI plays the key role in energy transfer. More energy will be transferred between high PFI particles that may cause high particle velocity. We also find that the fluctuation velocity square of particles presents the trend of first increasing then decreasing and finally increasing along the axial direction. The difference between PFIs is also elucidated by the total energy dissipation in each collisional energy level for a single particle. Results show that the single particle can endure intenser collision, more energy loss in anterior half segment than those in the second half of the grinding area. As mentioned above, the particle flow was analyzed in terms of particle flow intensity, particle velocity, collision energy, collision number, and so on. Some experimental results confirm the validity of the simulation. The simulation reflects the stable flow characteristics of particles in the grinding area and provides bases and references for further studying the product quality control and grinding equipment parameters optimization.