We propose a novel architecture for smart farming to enable the development of sustainable, energy-efficient, and reliable agricultural solutions. The proposed architecture utilizes Unmanned Aerial Vehicles (UAVs) as mobile platforms for data collection, fog computing, and wireless charging of smart farming devices. This architecture adopts a drone-as-a-service model to provide greater autonomy and resource utilization. We also propose a hierarchical processing scheme for big data analytics, thereby enhancing the latency, energy efficiency, and reliability of the data-driven smart farming technology. Moreover, we formally establish the parametric constraints imposed by wireless power transfer mechanisms on the topological characteristics of deployments. The provisioned agility of this architecture and its plug-and-play approach improve the scalability and cost-efficiency in deployment and management of smart farming systems.