Software requirement is become more important in recent because the development which witness in projects, badly executed requirements engineering steps can result in bad quality software and more cost for expensive maintenance. Manual classification of requirements is difficult, time-consuming, and expensive, especially in large projects and is written as a Software Requirements Specification (SRS) document. For this reason, automating software requirements classification helps in obtaining higher accuracy and saving time and effort. Most of researcher applied Intelligence techniques algorithms to avoid erroneous requirements and human intervention, as well as analyze, classify, and priority of requirements. In this paper illustrated modern of artificial techniques algorithm to classify RT approaches. It is surveyed that existing techniques like machine learning algorithms such as K-Nearest Neighbor (K-NN), decision tree (DT),.. etc.  Many other technical how ensemble learning and deep learning algorithm results in classification of RF. Researchers have proposed automated techniques to classify functional and non-functional requirements using several machine learning (ML) algorithms with a combination of different vector techniques. However, using the best method in classifying functional and non-functional requirements still needs clarification, and through many studies and research by researchers.