Diagnosis of skin disease in dark skin is extremely difficult, and this typically results in misdiagnosis or late diagnosis. This healthcare disparity impacts millions of individuals in the black community. Developing accurate and reliable Artificial Intelligence (AI) methods is crucial to address these challenges. In this paper, we applied three machine learning models to predict various skin conditions using dermoscopic images. The CRISP-DM data mining methodology was employed with three models: convolutional neural networks (CNNs), known for its strength in image recognition; support vector machines (SVMs), effective in distinguishing subtle differences in skin conditions; and decision trees, a simpler and interpretable model. These models were applied to predict skin disease types using the HAM10000 image dataset. Data imbalance was handled by applying SMOTE resampling method. For optimization the machine learning methods, fine tuning and principal component analysis (PCA) were applied. Accuracy was used as an evaluation measure. Among these models, SVM with the RBF kernel achieved the highest accuracy of 88.48%, followed by decision trees with 82.48% and CNN-ResNet50 with 77.28%. Comparisons with other methods on the same dataset demonstrated superior performance of the proposed models, particularly with the SVM-RBF model. These results suggest that the proposed models for skin disease detection have significant potential to benefit healthcare by offering more accurate and reliable diagnostic tools.