Abstract: The physical parameters of chili seeds are important inputs in the discrete element simulation. In order to improve the reliability and accuracy of the numerical simulation, the physical property parameters of capsicum seeds were measured experimentally, including triaxial size, density, Poisson’s ratio, friction coefficient, recovery coefficient, et al. Combined with the “Hertz Mindlin with JKR” viscosity model in EDEM, surface energy parameters were introduced to conduct the angle of repose test. The three most significant influencing parameters（chili seed-seed rolling friction coefficient, chili seed surface energy, and chili seed-steel plate rolling friction coefficient） were selected among the influencing factors by Plackett-Burman test, and the regression model of the three significant influencing parameters and the angle of repose was established to study the influence of the significant factors on the angle of repose under the interaction effect. The interval of significant parameters was divided into 5 levels by gradient average for climbing test to get the optimal range of values, and finally the response surface method（Box-Behnken） was applied to analyze the variance of the quadratic polynomial of the regression model in terms of the angle of repose, the relative error and the three important parameters, and the model was optimized with the minimum value of relative error to get the optimal combination: rolling friction coefficient between chili seeds was 0.75, JKR surface energy was 0.31 J/m~2, and rolling friction coefficient between chili seeds and steel plate was 0.60. The optimal combination of parameters obtained from the calibration test was used to conduct the simulated stacking test, and the average angle of repose was 26.71°, and the relative error was 4.61%.