Malignant melanoma is a very serious dermatological disease, which is rare but away more deadly skin cancer. Melanoma is the most dangerous form of skin cancer, while curable with early stages detection. Beside professional specialists that are capable of precisely identifying the disease, automated systems are also capable of recognizing disease which might save lives and reduce costs. Toward this goal, in this research, a classifier model based on support vector machine (SVM) with radial basis function kernel was fed with image features and class labels to predict the presence or absence of the malignancy in dermoscopy images. Based on this classifier, ABCD features (Asymmetry, Border, Color and Diameter) and texture features derived from the Haralick texture features calculated from Gray Level Co-occurrence Matrix were investigated to find the best features that could increase the accuracy of the diagnosis process. Eventually, results concluded 12 texture features with highest efficiency. In addition, results show that the new added texture features did improve the accuracy by a 9.6% than the common ABCD rule from 84% to 93.6%.