Package delivery services by deploying Unmanned Aerial Vehicles (UAVs) have received considerable attention due to their significant potential benefits. One of the scenarios making the concept more attractive is to deploy large numbers of UAVs to properly handle increasing demands for the service by the units. However, it is challenging because it increases the difficulty of UAV control as well as the complexity of the scheduling problems involved in the service. To tackle the issues, we propose a systematic approach that decomposes a service area into several disjoint zones. Within a zone, a single UAV at most can be operated. To verify the advantages and performance of the proposed approach, we first optimize the zoning, i.e. a service area decomposition, applying Genetic Algorithm (GA). We then evaluate the service level of the package deliveries by UAVs under the proposed approach using a simulation model.