To address the issue of insufficient accuracy in the discrete element simulation model for cotton stalk -residual film mixtures, this paper employs the Hertz-Mindlin model with the JKR contact model to calibrate the contact parameters. First, significant parameters affecting the angle of repose were identified using Plackett-Burman experiments. Subsequently, the optimal parameter range was determined in conjunction with the steepest climb test. A regression model was constructed using Box-Behnken experiments to optimize the parameters, yielding the optimal parameter combination: cotton stalk-residual film rolling friction coefficient of 0.46, residual film-residual film rolling friction coefficient of 0.31, residual film-steel rolling friction coefficient of 0.47, and residual film-residual film JKR surface energy of 0.41 J/m². Simulated pile-up tests were conducted on a mixture of cotton stalks and residual film using these optimal parameter combinations. The results showed that the error between the simulated and measured angle of repose was only 3.06%, validating the reliability of the parameters and providing a basis for research into the interaction characteristics between cotton stalks and residual film during film recovery.