Waste classification is a critical component of modern waste management systems. The fusion of cutting-edge technology with the ever-pressing issue of waste management holds great promise for a sustainable and cleaner future. Integration of transformer models in waste management systems can transform the landscape of solid waste classification and redefine the way of managing waste streams. This research paper contributes to this evolving field by presenting a comprehensive investigation into the utility of transformer models for solid waste classification. This paper addresses the challenge of solid waste classification using advanced deep learning models and presents a transformer-based architecture specifically tailored for this task. The importance of accurate waste classification in reducing environmental impact, optimizing recycling, and complying with regulations is emphasized. In experimental comparisons against established models, including Simple CNN, ResNet101, MobileNet, and DenseNet121, the proposed transformer model showcased superior performance, achieving a remarkable accuracy of 96.1%. This performance highlights the potential of self-attention mechanisms in effectively capturing intricate waste features. While other models performed well, the Transformer model's results underscore its significance in waste classification.