The proliferation of Internet of Things (IoT) networks demands secure, scalable, and efficient data management. Multi-Zone DAG Blockchain (MZ-DAG) has emerged as a lightweight solution, distributing data across zones to improve performance in IoT environments. However, existing MZ-DAG Blockchain designs lack robust privacy and adaptive trust management, exposing the system to evolving threats. This paper enhances MZ-DAG by integrating Zero Trust principles into its core design. First, we embed Homomorphic Encryption (HE) into the MZ-DAG data pipeline, enabling encrypted data processing throughout the blockchain lifecycle. This ensures that even fog nodes and validators handle only ciphertext, achieving Zero Trust Privacy Preservation, where no intermediate component can access plaintext data. Second, we enhance the Proof-of-Authentication (PoAh) consensus by introducing a Trust-based Node Selection mechanism modeled as a Markov Decision Process (MDP). Node trust evolves dynamically based on observed behavior, and trustworthiness is continuously re-evaluated using entropy-based uncertainty analysis. This adaptive selection prevents static trust assumptions and enforces continuous verification, aligning with Zero Trust’s “never trust, always verify” principle. Together, these improvements strengthen privacy, security, and resilience without compromising the low latency and lightweight design essential for IoT. This work establishes a comprehensive Zero Trust Multi-Zone DAG Blockchain model tailored for secure and efficient IoT data management, providing a foundation for future zero trust architectures in distributed IoT environments.