According to WHO estimates, over the next two decades with cancer, the count of patients diagnosed will double. It is possible to detect cancer in an early stage with the help of Artificial intelligence techniques. Detection of Melanoma at an early stage can be helpful in reducing the death rate to a great extent. So a Transfer Learning based EfficientNet-B2 model is used for the early detection of Melanoma. Efficient Net uses the concept of appropriate scaling of the model which efficiently balances network depth, width, and resolution. Balanced accuracy (BACC) is used as the main evaluation measure, where it is equivalent to the average sensitivity or recall. The proposed model scored a Balanced Accuracy 75.3%. The model achieves better classification accuracy than other state-of-the-art Deep Learning Models.