Today, the predominant issue in global health pertains to skin cancer, being the most prevalent type of cancer, encompassing variations like melanoma and basal cell carcinoma. Early detection is critical for effective treatment, yet traditional diagnostic methods often fail due to limitations in image quality and the complexity of visual differentiation. In this paper we examines novel methodologies elucidated with the objective of improving the dependability and accuracy of skin cancer detection using sophisticated computational techniques. The proposed approach employs sophisticated image pre-processing techniques to remove noise while preserving essential features. For this study we proposes an efficient deep learning approaches for optimal segmentation and classification of skin cancer with a focus on severity analysis will be implemented in python. Furthermore, the study includes a comprehensive analysis of the severity of the identified cancers to furnish a comprehensive comprehension of the advancement of the cancer and potential impact on the individual. Overall, the capacity of sophisticated deep learning methodologies to revolutionize the field of skin cancer diagnosis is noteworthy, providing a sturdy resolution that elevates the prompt identification, precise categorization, and assessment of severity, consequently enhancing the quality of patient care and results.