The rapid rise of electronic waste (commonly referred to as "e-waste") has become a world's growing challenge which should be managed by creative approaches. The number of e-waste produced is estimated to be 53.6 million metric tons in 2019. From this we can see that the seriousness of the issue direly calls upon taking the measures to prevent the environmental and public health risks associated with this expanding crisis [1]. Since a lot of the e-waste may contain hazardous materials such as mercury, lead and cadmium, which can impact the health and the environment if not treated properly, the mismanagement of it increases the problem [2]. In the case of e-waste, there is wide assortment of the electronic devices and components hence, it becomes difficult to classify them into their product categories properly. Sorting processes can't keep up with the pace of production waste as a result of being tedious, error-prone, and slow. This research employs deep learning approaches to segregate E-waste items using images for automated category. Utilizing contemporary models like VGG16, DenseNet121, InceptionV3, MobileNetV3, and ResNet50, the research designs classification systems that have these great attributes. Dataset building (training and assessment) become easy when an extensive dataset of 3000 images from 10 different types of equipment is developed. This research study helps to offer useful implications for managing current methods of electronic waste disposal and developing sustainable circular economies with quantitative analyzing of model performance factors that include accuracy, precision, F1- score, mean squared error (MSE), and mean absolute error (MAE).