This literature review focuses on the use of machine learning and deep learning techniques for skin disease detection and classification. The review covers studies from 2021 to 2023, sourced from the Scopus database, and includes 45 studies, with 32 using deep learning, 11 using machine learning, and 2 using a hybrid approach. The studies are evaluated based on models, datasets, and performance metrics. The review highlights challenges such as noise handling and disease symptom identification. Skin diseases are prevalent and can lead to severe conditions like skin cancer if not diagnosed early. Conventional diagnosis methods, relying heavily on expert dermatologists, are time-consuming and resource-intensive. Machine learning and deep learning offer promising solutions by leveraging large annotated datasets and advanced models. Feature extraction is a critical step in these techniques, and methods like MixPave (Mixed Prompts for few shots Attribute Value Extraction) are proposed to address overfitting and improve performance. The conventional diagnostic process involves histological analysis, dermoscopy, and visual inspection, but it is limited by the scarcity of experienced dermatologists and the time-consuming nature of the process. The review also discusses the advantages of computer-aided systems in providing cost-effective, comfortable, safe, and rapid diagnoses, making them valuable tools in healthcare.This literature review focuses on the use of machine learning and deep learning techniques for skin disease detection and classification. The review covers studies from 2021 to 2023, sourced from the Scopus database, and includes 45 studies, with 32 using deep learning, 11 using machine learning, and 2 using a hybrid approach. The studies are evaluated based on models, datasets, and performance metrics. The review highlights challenges such as noise handling and disease symptom identification. Skin diseases are prevalent and can lead to severe conditions like skin cancer if not diagnosed early. Conventional diagnosis methods, relying heavily on expert dermatologists, are time-consuming and resource-intensive. Machine learning and deep learning offer promising solutions by leveraging large annotated datasets and advanced models. Feature extraction is a critical step in these techniques, and methods like MixPave (Mixed Prompts for few shots Attribute Value Extraction) are proposed to address overfitting and improve performance. The conventional diagnostic process involves histological analysis, dermoscopy, and visual inspection, but it is limited by the scarcity of experienced dermatologists and the time-consuming nature of the process. The review also discusses the advantages of computer-aided systems in providing cost-effective, comfortable, safe, and rapid diagnoses, making them valuable tools in healthcare.