Abstract: ‍ ‍Reconfigurable intelligent surface （RIS） is viewed as a key technology for the next-generation 6G wireless communication system. By integrating a large number of low-cost passive reflecting elements on the antenna array， RIS can adjust the incident signals dynamically and reshape the wireless propagation environment， in which case an artificially controllable electromagnetic environment becomes a reality. Higher rate data traffic and network capacity， and the more intelligent network paradigms are the key requirements for 6G networks. The construction of smart radio environments can further help greatly improve the performance in terms of spectrum and energy efficiency， and support more intelligent designs in 6G networks. However， the introduction of RIS also significantly complicates the design of wireless communication networks， and the conventional techniques face much more new challenges and are even no longer applicable. In view of the characteristics and different application scenarios of RIS in 6G network， it is necessary to innovate and design new transmission techniques to fully unlock the potential of RIS deployments. Hence， we first introduced the basic concepts and the hardware framework of RIS， as well as its classification. In addition， the current various applications of RIS in wireless communication systems was further classified， and its five main application scenarios were summarized. Then， from the perspective of the techniques of the RIS-aided 6G network， we surveyed from the three aspects of channel modeling and channel estimation， hybrid beamforming， and the integration of artificial intelligence （AI） with RIS in 6G. The common RIS-related channel and path loss modeling methods were fully introduced. We then introduced the conventional techniques and the new ones， especially the AI based techniques， for channel estimation and hybrid beamforming in detail， together with some main challenges. Moreover， we focused on the design of RIS-aided systems with integrated sensing and communication （ISAC） and the deep integration of RIS and AI technology， as well as the construction of the RIS-assisted intelligent networks. Finally， the current application of RIS in practical systems and crucial problems to address were discussed and the potential future developments of the RIS technology were presented.