Abstract: Accurate and reliable contact and bonding parameters can greatly contribute to the simulations of rapeseed threshing during combined harvesting. In this study, a discrete element model of rapeseed stems was constructed to facilitate the calibration and optimization of the contact and bonding parameters using EDEM software. Initially, the contact model parameters were refined. A Plackett-Burman test was employed to identify the significant influencing factors on the bulk density angle of rapeseed stems. Multiple factors were determined to clarify the most substantial impact on the outcome. After that, the steepest ascent test was used to explore the best ranges for these parameters. The iterative approach was used to narrow down the parameter values for the best simulation. Furthermore, the Box-Behnken test was conducted to fine-tune the calibration of the contact parameters for the middle portion of the rapeseed stem using the response surface method. The parameters were also optimized to explore the interactions among the variables and their effects on the simulation. As such, mathematical equations were derived to integrate the experimental indicators into factors. An optimal set of parameter values was determined to validate their accuracy and reliability using additional experiments. The parameters of the bonding model were also determined as the most influential factors in the maximum bending failure force. A Plackett-Burman test was also applied to identify the key influencing parameters on the bonding characteristics of rapeseed stems. Subsequently, a quadratic regression orthogonal combination test was employed to derive a quadratic regression equation, particularly with the relationship between the significant factors and the bending failure force. The best combination of parameters was then identified for the bonding model. The most accurate simulation was achieved using the optimal parameters. The results reveal that three parameters—stem shear modulus, the static friction coefficient between the stem and the steel plate, and the static friction coefficient between stems were particularly influential in the contact model. The shear modulus was found to be 17.251 MPa, while the static friction coefficients were 0.353 and 0.383, respectively. In the optimized parameters, the relative error between the simulated and actual bulk density angles was minimized to 1.12%, indicating the high accuracy of the simulation. Furthermore, the unit area normal contact stiffness and the unit area tangential contact stiffness were identified as significant factors in the bonding model. The values of parameters were 1.89×10¹⁰ and 1.54×10¹⁰ N/m³, respectively. Once these optimized parameters were applied, the bending simulation closely matched the actual test, with a relative error of 0.36% for the maximum bending failure force. In conclusion, the rapeseed threshing was better simulated to accurately calibrate and optimize the contact and bonding parameters. The findings can also provide a robust representation of the mechanical properties of rapeseed stems when subjected to impact and bending. A great contribution can also be given to enhance the design and functionality of rapeseed combine harvester threshing devices. The structural parameters of harvesting equipment can also be optimized to improve the overall efficiency of rapeseed harvesting.