In the present scenario skin cancer is found highly risk in human beings. Many forms of skin cancer are affecting the human life. Among the form of skin cancer the unpredictable diseases is Melanoma cancer. Skin cancer the fatal form is primarily diagnosed visually leads to death, if not diagnosed in its early stage. It can be identified by tedious lab testing with more time and cost. There are vast numbers of computational techniques helpful to predict diseases. A challenging task in skin lesion classification is due to the smooth variation, in the appearance of skin lesions. Image processing techniques like segmentation is used in medical science to identify the region of significance.. This paper focuses Genetic algorithms by means of adaptive parameters (adaptive genetic algorithms, AGAs), an important and promising alternative to genetic algorithms. The extent for accurate solution and convergence speed is significantly measured by employing of crossover along with mutation from which genetic algorithms appear.