Solid waste management is a crucial environmental concern that calls for practical sorting methods to minimize trash and maximize recycling efforts. Machine learning algorithms have recently demonstrated considerable promise for same. This study focuses on three machine learning models: logistic regression, VGG-16, and DenseNet-169. First, trash objects are divided into categories based on the attributes using logistic regression, and then high-level features from trash photos are extracted using the deep convolutional neural network VGG-16 model. Lastly, the garbage sorting issue is addressed using the DenseNet-169 model. In DenseNet-169, the convolution of skips connections between layers improves the feature reuse and gradient flow of the architecture. It makes it possible to train the model quickly and make better use of the little database available, enhancing the model performance. An extensive dataset of garbage samples is compiled, annotated, and used for training and testing. Metrics such as accuracy, precision, recall, and f1 score are used to compare the models according to experiment findings. All three methods can sort the garbage effectively, each with distinct advantages and disadvantages, and accurately classify trash by capturing complex patterns and visual characteristics. This study shows that machine learning methods can automate solid waste sorting. The models offer precise and valuable ways to categorize waste materials, supporting effective waste management procedures and encouraging recycling initiatives.