Fungi play pivotal roles in ecosystems, agriculture, and industry, underscoring the importance of accurate classification for understanding their ecological significance and practical applications. Conventional classification techniques frequently require significant manual effort and can be subjective, leading to the creation of automated systems. This research project aims to revolutionize fungi classification by applying deep learning techniques, leveraging the comprehensive DeFungi dataset provided by UCI. The study meticulously designs and evaluates convolutional neural network (CNN) architectures, with a specific focus on ResNet, VGG, and InceptionV3 models. Through exhaustive experimentation, which includes meticulous data preprocessing, rigorous model training, and thorough evaluation, ResNet and VGG models consistently outperform InceptionV3 in terms of both accuracy and stability. The implementation of dropout regularization proves to be remarkably effective in preventing overfitting and enhancing the generalization capability of the models. Notably, the highest accuracy achieved in this study is an impressive 98.90%, attained by a ResNet model with a meticulously crafted architecture and dropout configuration. VGG models also demonstrate notable performance, achieving accuracies of up to 90.43%, showcasing their considerable potential for fungi classification tasks. Overall, this research contributes significantly to advancing fungi classification methodologies, providing practical and efficient solutions for researchers, practitioners, and industries alike. Furthermore, it underscores the importance of further exploration in hyperparameter optimization and model ensembling to further enhance classification performance and robustness across diverse fungi species.