Abstract: An accurate and reliable discrete element model of cotton terminal bud stem is often required to optimize the terminal bud removal device. However, it is still unclear on the interaction between the terminal bud removal device and the cotton terminal bud stem at the cotton topping stage. Taking the cotton terminal bud stem at the topping stage as the research object, this study aims to calibrate and optimize discrete element simulation parameters using EDEM software. Hertz-Mindlin non-slip model and Hertz-Mindlin bonding V2 model were used to construct the stem stacking angle model and the stem crushing model. The physical test was conducted to measure the stem stacking angle, maximum tensile failure force, and maximum shear failure force, thus taking as optimization objectives. Plackett-Burman, Steepest ascent and response surface tests were conducted for the experimental design. The parameters were then calibrated from the stem stacking angle and the bonding parameters of the stem crushing simulation model. Stacking angle and the tensile shear tests were optimized to verify the simulation. The parameters of cotton terminal bud stem after measurement were: The average diameter was 3.60 mm, the water content was 76.99%, and the density was 1 053 kg/m³. The stacking angle was measured by 30.58°; the maximum tensile and shear failure force were measured by 17.06 and 5.56 N, respectively. Plackett-Burman test analysis showed that the stem-stem static friction coefficient X₄, the stem-stem rolling friction coefficient X₅, and the stem-steel static friction coefficient X₇ shared the significant effects on the stem stacking angle. In fracture response, the unit-area normal contact stiffness X9 and the bonded disk scale X₁₃ also exerted significant influence on tensile and shear forces, indicating the sensitivity of failure loads to bonding stiffness and effective bonded area at particle contacts. Contact parameters were obtained after calibration: for stem–stem interactions, restitution 0.3, static friction 0.65, and rolling friction 0.53; for stem–steel interactions, restitution 0.39, static friction 0.73, and rolling friction 0.4. Calibrated bonding parameters were obtained as: unit-area normal stiffness 5.05×10¹⁰N/m³, unit-area shear stiffness 2.25×10¹⁰ N/m³, critical normal stress 5.8×10⁸ Pa, critical shear stress 3×10⁸ Pa, and bonded disk scale 0.66. The tensile shear test was conducted to verify the simulation, with the relative error of 3.58% for the tensile force between simulated (16.45 N) and the measured (17.06 N). The average shear force was 5.787 N after simulation, and the relative error of the 5.56 N was 4.08 % after measurement. The time-load curves of simulation and the physical test shared the same trend under loading, indicating that the parameter was calibrated to accurately reproduce the mechanical response of cotton terminal bud stem. Collectively, once the cotton apical-bud stem was subjected to systematic calibration and verification using discrete element method. High accuracy and reliability were obtained for granular repose and bonding fracture behavior. The findings can provide a quantitative basis to optimize the structural parameters in apical-bud removal devices.