Skin cancer is a common and potentially fatal condition, making early discovery critical to successful treatment. Recent machine learning (ML) approaches have shown quantifiable improvements in diagnostic accuracy, sensitivity, and specificity. Skin cancer arises from unrepaired DNA damage in skin cells, leading to genetic mutations or abnormalities. It often spreads slowly to other parts of the body, which makes early detection vital, as it’s more manageable in its initial stages. With skin cancer rates climbing, coupled with high mortality rates and significant medical expenses, early detection becomes increasingly important. Given the gravity of the situation, researchers have suggested some effectual approaches for early skin cancer detection. This study provides a systematic review of publicly available datasets and ML techniques used in skin cancer detection, detailing their capabilities, limitations, and applications. Challenges such as dataset bias, model interpretability, and ethical considerations are also discussed to offer a holistic perspective. Aiming to advance ML-driven skin cancer diagnostics, this work identifies gaps and suggests future directions for research.