The skin, as the body’s primary defense mechanism, shields internal organs from external threats. It faces a broad spectrum of challenges from microbes such as viruses, fungi, and bacteria, as well as the damaging effects of dust. Every year, a significant number of people worldwide suffer from various skin diseases, often contagious and rapidly spreading. The diverse nature of these skin conditions necessitates considerable expertise for accurate diagnosis and treatment by healthcare professionals. In recent years, there have been significant advancements in the field of diagnostic processes. Such development is the integration of deep learning models, specifically convolutional neural networks (CNNs) and NASNET. Our team has conducted a study that introduces a unique and effective method for classifying skin cancer, using a modified version of the HAM10000 dataset. Our model has shown exceptional performance when compared to the existing state-of-the-art models, all while utilizing fewer parameters. By leveraging advanced technologies like deep learning, the medical community aims to automate and improve the accuracy of skin disease diagnosis, ultimately enhancing patient care and outcomes. In comparison to the current deep learning models, this variant of convolutional neural network has achieved impressive accuracies of 98.7%, 97.04%, and 97.07% on its training, testing, and validation data sets. It has achieved these results using only 0.13 million parameters.