Remanufacturing is a key element of a circular economy, and used products are essential raw materials in a remanufacturing system. Arrangements must be made to collect these products from the end-users and deliver them to the remanufacturer. Research on optimising the routes of the vehicles employed in the collection process is limited. Because of the NP-hard nature of this reverse logistics problem, exact solution methods are not applicable, and researchers must resort to metaheuristics to find good solutions quickly. This study explores the use of three different versions of the Bees Algorithm for optimising collection vehicle routing, comparing them to other optimisation methods. The results show that the Bees Algorithm version that distributes bees for local search according to the Fibonacci sequence is the most effective in identifying near-optimal routes for collection vehicles. This work assumes that all the problem parameters are constant, which allows optimisation to be performed offline. Future research will investigate using digital twin technology to achieve online real-time optimisation.