Abstract: Abstract: Alfalfa is compressed into pellets, which can not only solve the problem of storage and transportation, but also maintain the nutrition of alfalfa. Therefore, alfalfa pellet has broad market application prospects. In the compression process, the compression piston overcomes the material deformation resistance, the friction between the materials and the friction between the materials and the inner wall of the die. Researching the effect of pressure on the material in the die is helpful to reveal the compressing mechanism and select the reasonable compression parameters. The discrete element simulation analysis provides an effective method for the stress research in the process of alfalfa straw densification. In the simulation analysis of alfalfa straw compression process with EDEM software, the accuracy of the input parameters has significant influences on the simulation results. The simulation parameters used in EDEM software for simulation of alfalfa straw compression were calibrated based on the combination of physical experiment and simulation optimization design in this study. The physical experiments of the contact parameters alfalfa straw-alfalfa straw and alfalfa straw-45 steel were carried out with test equipments such as inclinometer and high-speed camera, and the collision recovery coefficient of alfalfa straw-alfalfa straw ranged from 0.1 to 0.3 while alfalfa straw-45 steel was 0.1-0.6, static friction coefficient of alfalfa straw-alfalfa straw was 0.3-0.6 while alfalfa straw-45 steel was 0.2-0.8, and rolling friction coefficient of alfalfa straw-alfalfa straw and alfalfa straw--45 steel were 0-0.3, respectively. Based on the physical experiment results of contact parameters, the Plackett-Burman test was used to screen the initial parameters. The results of variance analysis showed that the static friction coefficient of alfalfa straw-alfalfa straw, rolling friction coefficient of alfalfa straw-alfalfa straw, static friction coefficient of alfalfa straw -45 steel had a significant effect on the simulation repose angle. Taking the relative error value of repose angle as the evaluation index, the steepest climbing test was carried out to optimize the range of significant parameters. Based on the Box-Behnken test, the second-order regression model of repose angle and significance parameters was established, Taking the 38.88° repose angle obtained from the physical experiment as the target value, the significance parameters were optimized and the optimal combination of parameters was obtained: alfalfa straw-alfalfa straw static friction coefficient 0.45, alfalfa straw-alfalfa straw rolling friction coefficient 0.08, and alfalfa straw -45 steel static friction coefficient 0.54. Finally, the parameters were used in the simulation experiments and repeated three times. The simulated repose angles of alfalfa straw were 39.62°, 37.83°, 38.69°, respectively. The results of T-test (P = 0.778> 0.05) with Origin 2018 software indicated that there was no significant difference in repose angle between the simulation and the physical test, which verified the reliability of the optimal parameter combination. The results showed that it was feasible to calibrate the contact parameters of alfalfa straw with the above optimization experiments. The calibration method can provide reference for the discrete element parameter calibration of other materials and the parameters obtained by the calibration method in this study can provide a parameter basis for the further study on the forming mechanism of the material block and the law of stress transfer in the material layer by using the discrete element method.