With the rapid development of big data and the Internet of Things technology, people have easier access to data, which leads to different perspectives when observing data. As a result, multi-view data has emerged, which can effectively demonstrate different manifestations of a sample, but is often more complex to process. Multi-view clustering has received increasing attention in recent years, utilizing the consistency and complementarity among multiple views to conduct more effective digital representation, in order to obtain information that cannot be obtained from a single view and improve clustering effectiveness. This article provides a classification and detailed discussion of existing multi-view clustering methods, and summarizes current challenges and future development trends.