Abstract: In order to achieve the synergistic optimization of two key process parameters—heating temperature and mixing time—during the mixing process of recycled asphalt mixture (RAP), this study constructed a multi-objective comprehensive scoring model based on Random Forest (RF) - Response Surface Methodology (RSM) after conducting 29 experimental trials, enabling rapid optimization of the process parameters. The experiments were designed using a uniform-random mixture sampling approach, and five key performance indicators—voids, stability, residual stability after immersion, dynamic stability, and freeze-thaw splitting strength ratio—were measured according to JTG E20-2011 to establish a comprehensive scoring function. An RF regression model was used to create a nonlinear response surface coupling model between process parameters and the comprehensive score, and finally, the optimal heating temperature and mixing time were determined through grid search. The results indicate that the comprehensive scores of the experimental samples ranged from 46.3 to 72.1 points, with heating temperature having a more significant impact on the score. The RF-RSM model has a coefficient of determination (R²) of 0.948 and a root mean square error (RMSE) of 1.78. The optimal process parameters were a heating temperature of 125°C and a mixing time of 115 seconds. This model demonstrates strong portability and engineering application value, effectively improving RAP process optimization efficiency and providing a data-driven optimization method for green RAP production.