Abstract: Abstract: The pneumatic metering device has the advantages of high qualified index, less mechanical damage to seeds and good adaptability for high-speed operation. But traditional calculation method is not capable of analyzing the law of seeds motion because of the complex stress state. Discrete element method (DEM) is the key method to solve the above problem. However, physical characteristics parameters of particles, mainly consisting of static friction coefficient and rolling friction coefficient between corn seeds, are key factors for reliable simulation results in the DEM. Due to the difference between corn seeds and simulation particles in surface shape and size, the actual physical parameters of corn seeds are not suitable for the EDEM (enhanced discrete element method). Particles models established in the EDEM are rougher than corn seeds, which leads to the distortion of simulation results. Therefore calibration works for particle parameters are essential. In this paper, a new method was proposed for actively seeking the target parameters based on regression analysis. It was estimated that repose angle was commonly influenced by static friction coefficient and rolling friction coefficient between corn seeds. All factors simulation experiments were conducted to obtain the regression equations, of which the tangent value of repose angle at different contact material was separately taken as the dependent variables, and the static friction coefficient and rolling friction coefficient between corn seeds were taken as independent variables. In order to decrease the measuring error, MATLAB software was used for images acquisition and analysis. The variance analysis showed that the static friction coefficient and rolling friction coefficient between corn seeds had significant influence on the repose angle and natural heaping angle separately at significance level of 5%. The tangent value of repose angle had an increase trend with the increase of static friction coefficient and rolling friction coefficient between corn seeds. For the first test with organic plate, it had a bigger increase when the static friction coefficient was at 0-0.2 and then increased slowly, at this time which was mainly influenced by the rolling friction coefficient. For the second test with aluminum plate, it also increased with the increase of static friction coefficient and rolling friction coefficient between corn seeds in a consistent state. By the simulation tests, 2 regression equations were obtained. For solving the independent variables, the actual tangent values of repose angle with different contact material were measured by the actual tests and taken as the dependent variables. By this way, the static friction coefficient of 0.0338 and the rolling friction coefficient of 0.0021 between corn seeds were obtained. Verification tests with calibrated physical parameters were carried out. Results showed that the tangent values of repose angle were consistent with the actual tests. The simulation model of corn seeds was composed of multiple small spherical particles. There were sunk areas between adjacent seeds, so the contact area between corn seeds increased and meanwhile the friction force was increased. Corresponding relation between actual seeds and simulation model was found by this calibration method. The boundary of repose angle was extracted to compare the simulation and actual test results, which verified that the calibrated physical parameters were applicable for the EDEM and the method was feasible.