This paper shows a CVRPTW (Capacited Vehicle Routing Problem with Time Windows) problem applied to collecting donations from a food bank. The main objective is to minimize the costs and operating times that are affected by the fact of being subject to a dynamic variation in the programming of donors to visit. This problem is solved by means of a mixed integer linear programming model that allows defining the best route system, using information extracted from the Google Maps API for calculating the distance matrix. The model plans the sequence of visits that the vehicles must make, which includes the main, donor and secondary nodes, which are common points between the routes through which the vehicle must pass.