In the last few decades, the generation of waste has exponentially increased throughout the world, which has subsequently resulted in the pressing need to recycle waste. The challenging issue in recycling waste lies in how the waste is being sorted. Traditional waste sorting methods are labor-intensive and limited in scalability, which increases the necessity of AI-driven approaches. In this research work, multiple machine learning models are trained using TrashNet image datasets to efficiently classify waste materials. To conduct a detailed analysis of the waste classification, the present work trains the traditional machine learning models such as SVM, Random Forest, KNN, Naïve Bayes, and Decision Trees. Also, deep learning such as (MobileNetV2, DenseNet201, ResNet101, Xception, and InceptionV3) are trained by applying normalization and noise reduction to optimize accuracy. To evaluate the performance of the proposed model, accuracy, precision, recall and F1-Score of each model is assessed individually. Upon careful consideration and evaluation of these trained model, DenseNet201 attains the maximum accuracy of 0.73 in deep learning models and Random Forest attains the maximum accuracy of 0.68 in traditional deep learning models. Additionally, this research work highlights the ability of deep learning models in identifying complicated waste patterns.