Skin diseases are among the most common health conditions, and early detection is crucial for effective treatment and prevention of serious complications. However, traditional diagnostic methods require expert consultation and specialized equipment, which may not be easily accessible to everyone. To address this challenge, this paper presents an AI-based system for early screening and detection of skin diseases using deep learning techniques. The proposed system utilizes a Convolutional Neural Network (CNN) model to classify skin images into three categories: melanoma, benign, and normal. The model is trained on a publicly available dataset and incorporates preprocessing techniques such as image resizing and normalization to improve performance. A user-friendly web interface is developed using Streamlit to enable real-time image input and prediction. The system provides classification results along with confidence scores, enhancing reliability and usability. Experimental results demonstrate that the proposed approach achieves satisfactory accuracy for clear and well-defined images. This system serves as an assistive tool for preliminary diagnosis and promotes early awareness, while not replacing professional medical evaluation.