Skin cancer remains one of the most fatal types of cancer in today's world. The high mortality rate and cost of medical services can be brought down drastically if the disease is detected and identified in earlier stages. The use of technology to combat such diseases is now essential and AI remains the frontrunner in such technology and use of Deep learning—CNN can be used to improve such technologies. In this chapter, we propose an average ensemble model based on four CNN models(InceptionV3, EfficientNetB7, DenseNet201 and Xception) to classify various types of skin cancer from the International Skin Imaging Collaboration(ISIC), 2020 dataset which consists of 9 classes Actinic keratosis, basal cell carcinoma, dermatofibroma, pigmented benign keratosis, melanoma, nevus, seborrheic keratosis, squamous cell carcinoma, vascular lesion. Our proposed weighted average ensemble learning model stands sound owing to its simplicity in execution, low cost and high accuracy(https://www.w3.org/1998/Math/MathML" display="inline"> 90.63 %), Precision(https://www.w3.org/1998/Math/MathML" display="inline"> 94.11 %) and Recall(https://www.w3.org/1998/Math/MathML" display="inline"> 85.94 %).