In modern society, environmental pollution and climate change are considered the main problems affecting to increase in cancer cases. One of these, skin cancer occurs when there is an overgrowth of abnormal cells in the skin. Additionally, Skin cancer can develop in areas faced with UV (Ultraviolet) radiation. But it can also form in areas that rarely see the light. Thus, many experiments in both medicine and computer areas were realized to diagnose and treat this illness. This study introduces SMSC (Sampling in MobileNet for Skin Classification), a framework that leverages a fine-tuned MobileNet model and advanced sampling techniques to address class imbalance in the HAM10000 dataset. SMSC achieved remarkable results, with a validation accuracy of 96.93% and a test accuracy of 95.61% for classifying seven skin cancer types. Additionally, for binary classification between benign and malignant lesions, the model reached an average validation accuracy of 99.41% and a test accuracy of 98.92%. SHapley Additive exPlanations (SHAP) was employed to provide interpretability by explaining the model’s decisions at the pixel level.