Artificial Intelligence (AI) is now used in many medical tasks like detecting diseases, analyzing X-rays, and predicting patient risks. But most AI models work like black boxes, so doctors cannot understand how decisions are made. Explainable AI (XAI) helps solve this problem by showing which image regions or patient features influenced the result. In this review, we study ten research papers that use XAI methods such as Grad-CAM, LIME, SHAP, and Attention for medical diagnosis. These techniques make AI more transparent and help doctors trust the predictions. The results show that XAI improves accuracy, safety, and understanding in both imaging and clinical data. However, issues like unstable explanations and limited clinical testing still exist. The paper also discusses future scope like multimodal XAI, causal explanations, and real-time hospital use. Overall, XAI is an important step toward safe and trustworthy medical AI.