The Industrial Internet of Things (IIoT) has transformed industrial ecosystems by enabling real-time monitoring, automation, and data-driven decision-making. Deep learning techniques have emerged as powerful tools for predictive analytics, supporting applications such as anomaly detection, fault diagnosis, and predictive maintenance. However, centralized deep learning approaches introduce significant security and privacy risks, including data leakage, adversarial attacks, and model poisoning. This research proposes a secure hybrid deep learning framework integrating CNN-LSTM with ANFIS, along with federated learning, blockchain, and differential privacy to ensure secure, privacy-preserving, and explainable predictive analytics in IIoT environments. The framework enhances prediction accuracy while maintaining data confidentiality, robustness, and real-time performance.