E-scooters are gaining popularity for short-distance travel, but their recharging presents challenges. To reduce their downtime, we propose a Hybrid K-Means/Particle Swarm Optimisation (PSO) approach, optimizing charging routes using machine learning and meta-heuristics. The research in this paper attempts to determine if a combination of a meta-heuristic such as PSO and a machine learning algorithm for clustering such as K-Means, would be effective at solving the vehicle routing problem for e-scooters. We compared this method with other algorithms and found that Tabu Search excelled in over 95% of tests. While Hybrid K-Means/PSO led in only approximately 52% of scenarios, it was also the only one to provide an output that surpassed Tabu Search in one of the scenarios. The core difference in efficiency is due to traditional meta-heuristic methods providing routes that while optimal, may also travel from locations relatively far from each other, while Hybrid K-Means/PSO will provide routes between locations that are clustered and in local groups. This results in Hybrid K-Means/PSO being slightly less efficient but may be more practical for charging personnel as they can operate in designated areas close to each other rather than a more optimal route with nodes further apart. This research underscores the effectiveness of Tabu Search and the potential of our Hybrid K-Means/PSO approach for optimizing e-scooter charging routes.INDEX TERMS e-scooter rechargeable; hybrid optimization k-means/particle swarm; tabu search; guided local search; simulated annealing.