In the rapidly developing field of the Internet of Things, efficient data dissemination and storage play a critical role, especially in resource-constrained and intermittently associated Wireless Sensor Networks. This paper proposes an agent-based, A dispersed data replication technique with minimal complexity that aims to improve data privacy, availability, optimizing energy consumption, and ensuring balanced memory utilization in FogInternet of Things environments. By leveraging static and mobile agents across the sensing, fog, and cloud layers, the system intelligently monitors node parameters such as memory availability, energy levels, and update freshness to make dynamic decisions for data replication. The proposed method ensures minimal data loss and prolonged network lifetime through the selection of optimal neighbor nodes based on a greedy agent-based heuristic. Also, this method proposes data privacy by using federated learning algorithm, which helps in processing data faster and efficient. Simulation outcomes reveal that the approach significantly outperforms traditional methods in terms of data dissemination efficiency, computation overhead, energy consumption, throughput and data processing time.