During residual film recovery operations, the mixing of residual film with soil and the absence of accurate interaction parameters significantly hinder the optimization of key machinery components. To tackle this issue, this study develops a discrete element simulation model based on the Hertz–Mindlin with JKR (Johnson-Kendall-Roberts) contact model, using the residual film-soil mixture as the simulated material. The contact parameters were calibrated through a combination of physical experiments and EDEM simulations. Using Design-Expert software, angle of repose tests and Plackett–Burman experiments were carried out to identify the parameters that have a significant influence on the angle of repose. The optimal set of parameters was determined through steepest ascent and Box–Behnken design (BBD) experiments: residual film–soil rolling friction coefficient 0.61, soil–soil rolling friction coefficient 0.60, soil–steel static friction coefficient 0.68, and soil JKR surface energy 0.46 J/m². Validation results indicate relative errors of 4.58% for the angle of repose and 2.58% for the uniaxial compressive strength between simulated and measured values. These outcomes offer a useful reference for optimizing residual film recovery equipment and for modeling soil–residual film composite materials.

During the harvesting process, rigid materials are prone to causing damage to the cotton stalks, which will increase the risk of stalk breakage. A cotton stalk pulling component that blends stiff and flexible materials was devised to lower the breaking rate. The cotton stalk pulling component was made up of rollers and flexible belts that pull the stalks using clamping force and the forward speed of the tractor. The influence of various factors in the equipment on the harvesting effect of cotton stalks were analyzed through response surface experiments, and a multiple quadratic regression response surface model with missing pulling rate and breakage rate as response values was established. The significant of influencing factors on the breaking rate of cotton stalks are in a descending order as: the angle of cotton stalk pulling, tractor’s forward speed, and the clamping speed of the cotton stalk component. The working parameters of the wheel-belt type cotton stalk pulling machine have been optimized using the response surface combination experimental method, and the optimal parameter combination was obtained as: tractor forward speed of 4.5 km/h, cotton stalk pulling angle of 60°, and clamping speed of the cotton stalk pulling component of 349 r/min. The results of validation experiments showed that the missing pulling rate of cotton stalks was 5.06% and the breakage rate was 13.12%, indicating a good harvesting effect of the cotton stalks. The model was reasonable and the performance parameters could meet the relevant inspection requirements. The results can provide a reference for further research on the technology of flexible cotton stalk pulling.