Abstract: Discrete element method has been widely used to design and optimize the orderly harvesting equipment for rape shoots. This study aims to improve the accuracy and reliability of discrete element parameters for the clamping section stems during rape shoot harvesting. The type of double low cabbage "Nongda No.1" rapeseed was used as the research object and then transplanted on the double row ridges. Some parameters were determined in the period of mechanized harvest, including the intrinsic, surface contact, and crushing mechanical parameters. The stacking and crushing models of simulation and calibration were established for the clamping section of stems using Hertz-Mindlin non-slip and Hertz-Mindlin with bonding model on the EDEM software. The parameters were gradually adjusted to approximate the simulation with the physical test values. The optimal model was obtained for the clamping section of stems after the stacking simulation and calibration. The contact parameters were also calibrated and optimized on the surface of the clamping section of stems. There were small relative errors in the rest angle between the simulation and physical test. The relative errors of the maximum axial compression force and the maximum bending force between the actual physical tests and simulation were taken as the targets to optimize the model. The calibrated parameters of surface contacts were applied to the bonding model. The calibration and optimization of the bonding parameters were then realized for the clamping section of stems. Finally, the calibrated parameters were verified using crushing simulation and mechanical tests of radial compression and shear, internal core and skin tension of the clamping section of stems, and orderly harvesting tests. The results showed that the diameters of the whole stem and the inner core of the mechanized harvestable-stage clamped stem were found to be 13.52 and 12.08 mm, respectively, with an epidermis thickness of 0.72 mm. The densities of the inner core and epidermis were 0.93 and 1.40 g/cm³, respectively. The collision recovery coefficients were 0.35, 0.30, and 0.27, respectively, for the epidermis-epidermis, epidermis-inner core, and inner core-inner core stem collisions. The collision recovery coefficients for the epidermis-soft PVC and inner core-soft PVC were 0.44 and 0.34, respectively. The average angle of repose was 34.95° for the clamped stem samples. The maximum axial tensile forces were 22.60 and 10.95 N, respectively, for the inner core and epidermis. The average maximum bending force and average deflection of the clamped stem were 31.30 N and 9.89 mm, respectively. The average maximum axial compressive force and average compression were 101.1 N and 2.70 mm, respectively. The average maximum radial compressive force and average compression were 46.50 N and 2.13 mm, respectively. The maximum shear force was 40.50 N. The static friction coefficient between the epidermis and soft PVC was found to be 0.548, while the dynamic friction coefficient was 0.449. The static friction coefficient between the epidermis and epidermis was 0.258, while the dynamic friction coefficient was 0.017. Significant parameters were obtained: the normal/tangential stiffness of inner core-inner core was 4.14×10⁸ N/m, the tangential stiffness of epidermis-inner core was 1.16×10¹⁰ N/m, and the tangential stiffness of epidermis-epidermis was 3.01×10⁹ N/m. The mechanical experiments were carried out on the radial compression and shear, as well as the tensile deformation of the inner core and epidermis. There were relative errors of less than 5% in the radial compression force and shear force between the simulation and actual test, as well as the tensile force between the inner core and epidermis. The "time-load" curve between the simulation and the actual test had a consistent overall trend, with a better linear relationship in the ascending stage. The authenticity and reliability of the contact and bonding parameters were further verified after calibration, indicating the correct and feasible model. At the same time, the actual clamping harvesting was carried out on the cropping, clamping cutting, and transportation operations. The general adaptability of the segment stem model was further verified after the clamping harvesting. There were relative errors of less than 7.0% in the low-, medium-, and high-speed gears between simulation and physical experiments, compared with the coupled RecurDyn-EDEMing simulation. It is feasible to apply the discrete element method to the ordered clamping harvesting in the period of mechanized suitable harvesting for the rape shoots. The calibrated model can provide theoretical guidance on the machinery coupling interactions.