Skin cancer is a peculiar growth of skin cells; often develops on skin exposed to the sun. Skin cancer evolves predominantly on areas of sun-exhibited skin includes scalp, face, lips, ears, neck, chest, arm, hands and the legs in human. But it may appear on the areas that seldom see the sunshine of day like palms, at a lower place fingernails or toenails and reproductive organ space. However, it is curable if it gets detected at an early stage. To minimize the diagnostic error caused by the complexity of visual interpretation and subjectivity, it is important to develop a technology like computerized image analysis. This article presents a systematic approach for the classification of skin lesions in dermoscopic images. For detecting the melanoma skin cancer, the Otsu algorithm is applied to segment the lesion from the entire image. In the proposed method, feature extraction is performed by underlying shape, color, texture using GLCM (Gray-Level Co-Occurrence Matrix) features which are applied with support vector machine and Bayesian classifiers. While comparing, the Bayesian classifier achieved an overall classification accuracy of 90.44% on a dataset of 4151 images which is better than the existing method.