Dermatological infections are among the most frequent ailments on the planet. Despite their high prevalence, assessment is challenging due to the intricacies of complexions, colour, and hair appearance. Skin diseases are a significant public health hazard worldwide. These infections become dangerous when they reach the invasive stage. The medical community faces a massive challenge in recognizing and monitoring skin diseases. The number of people suffering from skin diseases is expanding faster because of increased pollution and poor nutrition. People frequently underestimate the consequences of skin diseases in their early stages. The present treatment process involves examining a biopsy procedure, and doctors manually administer drugs. This paper offers a hybrid method to circumvent human examination and provide optimistic timely results. In-depth research methods and deep learning approaches may aid in creating capable frameworks for categorizing various forms of skin diseases. To begin identifying skin disorders, it is necessary to distinguish between the skin and nonskin tissue. This research suggested a computerized classification method for skin diseases based on Deep Convolutional Neural Networks (DCNN) and Binary Butterfly Optimization Algorithm (BBOA). This approach aims to maximize the accuracy of skin disease prognosis while retaining a high level of effectiveness in collecting stateful knowledge for precise predictions.