Optimizing delivery routes for efficient goods distribution is a critical challenge in logistics. The Vehicle Capacity Routing Problem (CVRP) is a mathematical model used to address this challenge. What remains to be done is to find the appropriate algorithm and method to lead us to the best possible solution given the constraints. This article takes an open-source referenced memetic algorithm (MA) used to solve CVRP to study the results of other possible configurations. We propose a comparative study of the results of different order crossover (OX) techniques, in terms of solution obtained and computation time, for a set of CVRP reference instances. This analysis will allow us to identify the most efficient variant of OX to solve CVRP.