Human skin is a remarkable structure. It regularly experienced several recognized and unidentified diseases. Therefore, the most complex and ambiguous area of research is the diagnosis of skin disorders in humans. The manual diagnosis of skin illnesses by medical professionals is subjective and time-consuming. Because of this, medical professionals and patients both need automated skin disease prediction, which expedites the treatment strategy. As a result, this research effectively suggested an innovative machine-learning strategy to address this issue. Initially, this study gathered an image dataset. After that, we de-noise the images using a bilateral filter (BF). Skin diseases are successfully identified by a novel Static Tree Twin Support Vector Machine (STTSVM) approach that makes use of the retrieved attributes. A number of current techniques are used to validate the suggested method's performance. The recommended technique outperforms existing approaches by effectively as demonstrated by experimental results.