We study a zoning approach that divides a service area into multiple zones where limited numbers of unmanned aerial vehicles (UAVs) are deployed to a zone. A solution algorithm using clustering and simulation techniques is developed to optimize the safety of a UAV fleet and service quality to demands.