Healthcare activities in hospitals generate numerous types of post-use waste materials that can be classified as hazardous. This study proposes an Artificial Intelligence (AI) and Internet of Things (IoT) integrated framework for secure and efficient hazardous waste management in hospitals. Smart bins with IoT-enabled locks ensure waste collection, while Convolutional Neural Network (CNN) and Adaptive Neuro Fuzzy Inference System (ANFIS) improve detection and classification accuracy. A kinematic waste sorting mechanism is proposed to manage space constraints in hospitals. Deep Reinforcement Learning optimises disinfection scheduling and waste storage, and Federated Learning ensures secure decentralised data handling. Preliminary models demonstrate significant improvements in classification accuracy, reduced manual intervention, and compliance with safety policies. This theoretical framework provides a scalable solution for hazardous waste management in healthcare and other industries, with a small-scale experiment that validates AI models.