Skin disease classification plays a critical role in dermatological diagnostics, impacting millions worldwide with various conditions. Traditionally, this classification process has relied on manual methods, which are prone to errors and consume significant time. However, the emergence of deep learning, particularly Convolutional Neural Networks (CNNs), offers a promising solution by automatically extracting hierarchical features from image data. The AlexNet architecture, chosen for its balance between model size and performance, proves particularly beneficial in resource-constrained environments like healthcare settings. The manual classification of skin diseases has long been recognized as inefficient and error-prone. Therefore, there is a growing need for more efficient solutions. Leveraging CNNs, specifically the AlexNet architecture, we achieved outstanding results, with a training accuracy of 99% and a robust testing accuracy of 96%. This highlights the potential of deep learning techniques in automating the identification and classification of diverse skin conditions.