In this research, we use supervised machine learning to enhance a metaheuristic algorithm designed for solving the Capacitated Vehicle Routing Problem (CVRP). The CVRP, a combinatorial optimization problem, involves determining the min-cost routes from a depot node for a fleet of capacitated vehicles to fulfill the demands of various customer nodes. Traditionally, the algorithms solving this problem start from a new solution each time, even when dealing with similar problem types. However, utilizing historical data could offer valuable insights for achieving more efficient and effective solutions. Moreover, the integration of machine learning (ML) holds the potential for real-time problem learning, guiding the algorithm toward more efficient problem-solving. Therefore, the objectives of this research is to establish an effective learning process combined with a robust optimization algorithm for solving problems with increased efficiency.