The waste hierarchy, epitomized by the 3Rs: reduce, reuse, and recycle, has long been a guiding principle in waste management. However, coastal regions face an escalating crisis due to inadequate waste management practices, which pose significant threats to marine ecosystems. Contributing factors include the lack of recycling awareness, population growth, and the increased production of waste materials such as plastics, aluminum cans, and paper. These challenges necessitate an automated, efficient approach to garbage detection to mitigate environmental impact. This research aims to enhance the speed and efficiency of garbage detection in coastal regions through the application of advanced machine learning techniques. Specifically, we employ the Single Shot MultiBox Detector (SSD), a state-of-the-art object detection algorithm known for its balance of accuracy and speed, outperforming methods like YOLO (You Only Look Once) in certain contexts. The SSD framework is integrated with VGG16, a deep convolutional neural network that serves as the backbone, facilitating robust feature extraction. Our dataset encompasses various classes of waste materials, including paper, plastic, and cardboard, which are critical for comprehensive detection. Through meticulous training and evaluation, the SSD Multibox model with VGG16 backbone achieved an accuracy of 74%, demonstrating its viability for real-time waste detection applications. The automated detection system proposed in this study not only accelerates the identification process but also contributes to more effective waste management strategies in coastal areas. This research underscores the potential of leveraging deep learning models to address environmental challenges and highlights the need for continued innovation in the field of waste management.