Skin conditions are common and have grew in recent years, according to the abstract. Skin conditions must be promptly and accurately diagnosed in order to improve patient outcomes and provide appropriate care. The subjective and time-consuming manual diagnosis procedures dermatologists use necessitate the use of automated techniques to improve accuracy and efficiency. This work aims to make use of the 1,000-item. We applied Deep Convolution Neural Networks that have been successful with the ImageNet dataset at classifying seven different types of skin lesions.With the MobileNet Model we produced results that are 95% accurate. By this model we are able to achieve more accuracy than previous one. This model makes use of Keras and TensorFlow, both of which are implemented in Python 3.8.0. This model precisely finds seven sorts of skin diseases with an exactness of 95%. The results of the experiments confirm the intuition that pre-trained models advance highlights, and the geographic properties of DCNNs aid in learning highlights for a dataset of skin lesions dermatoscopic images, which originates from something altogether different.