Technological innovations in the Internet of Things (IoT) enable the acquisition, distribution, and interpretation of massive data, facilitating the development of numerous versatile and responsive applications in real-world contexts. However, IoT faces challenges in its centralization, privacy preservation, latency, and security. This paper uses prominent decentralized technologies like blockchain and federated learning to provide Quality-of-Services on 6G mobile networks to preserve privacy. In order to protect user privacy, user equipment (UE) at the distributed federated layer is trained using a machine learning model that masks sensitive data from service providers or masqueraders. Consequently, it empowers UE to converge on a global model by executing local gradient updates with network automation. We leverage blockchain as a decentralized sharing model over the federated layer to allow wireless networks to offload computational load in rendering real-time IoT data with an accuracy of 98%. A novel dynamic authentication framework is proposed to enable UE to interact with various base stations in the 6G wireless network with an average computation, aggregation and response time of 105 ms, 3.84 ms and 45.2 s respectively. The Distributed Hash Table at the core Blockchain layer offers decentralized secure storage with an improved throughput of 1890 TPS.