Abstract: Abstract: Tenderness is one of the most important factors influencing the quality of beef. Traditional evaluation methods have some disadvantages and limitations more or less. In order to predict beef tenderness accurately, conveniently and objectively, in this research, the discrete element method was used to establish the beef chewing model. Beef from the mid-region of longissimus dorsi (LD) was collected from 50 cattle as the samples, in which 30 cattle were used for structuring the beef chewing model, and 20 cattle were prepared for verifying the accuracy. The age of cattle (400-550 kg) was from 30 to 36 months, and the cattle were fattened for more than 6 months. After starving for 24 h, the live cattle were weighed, showered, stunned, killed, and bled blood. The 4 limbs and head of each animal were cut off, and the body of cattle was split into halves, cooled at 4℃for 24 h, and then the carcasses were divided. Each piece of beef was cut into 10 mm × 10 mm × 10 mm sample, but the inter-muscular fat, connective tissues and tendon were deleted. The samples were placed into plastic bags individually in a 75-80℃water bath, and cooked for 15 min until the internal temperature of beef sample reached 70℃. The samples were divided into 3 groups so as to carry out the experiments in triplicate after the samples were cooled to room temperature (20℃). Shear modulus and normal stiffness were detected by Brookfield CT3 texture analyzer (Brookfield Engineering Laboratories, INC. Middleboro Massachusetts, USA). With a two-cycle texture profile analysis (TPA) model (a compression model for normal stiffness) and a TA44 probe (cylinder diameter=4 mm), the size of testing surface of each sample was 10 mm ×10 mm × 10 mm (for normal stiffness). The related parameters settings were: test speed of 0.5 mm/s and deformation quantity of 2.5 mm for shear modulus detection, and test speed of 0.5 mm/s and preload of 2 N for detecting normal stiffness. In addition, density, restitution coefficient, friction coefficient and other parameters were also gained by the experiments to establish the beef chewing model. This chewing model made use of the particle body force in the Hertz-Mindlin with bonding model to replace the beef with micro particles which was the minimum unit in simulation. During the chewing process, the particle cluster would be not broken until the normal stress and tangential stress between random 2 micro particles exceeded the maximum limit values. The enhanced discrete element method (EDEM) recorded the exchange of force and beef particles of samples during the chewing cycles. Chewing simulation continued for 4 cycles and got the average shear force to judge the level of tenderness. Then the sensory evaluation and texture analysis were used to verify the results of simulation and compare their accuracy. Ten healthy and dentally tidy adults were chosen to evaluators who had the age from 20 to 25 years old and without thirst or hunger. Each evaluator chewed the samples and estimated the level. Accuracy of broken process and stress change during beef chewing by the EDEM software achieved 90% compared with the sensory evaluation method. On the other hand, the texture analysis showed that the parameters were test speed of 0.5 mm/s, moving distance of 2.5 mm and preload of 2 N. Compared with texture analyzer's results, the accuracy achieved 90% as well. The results prove that the discrete element method is a new efficient method for beef tenderness quality inspection.