Abstract: ‍ ‍In recent years， people have been developing reconfigurable intelligent surface （RIS）， which is a low-cost structure composed of many passive reflectors. It can freely change the reflection coefficient of its each element and reconfigure the wireless transmission environment to change the propagation of the incident signal. Due to several novel and unique advantages， RIS has a wide range of application scenarios， and research on it in physical layer security flourishes. At the same time， eavesdropping channel state information in the physical layer security has always been a focus of people while the actual communication scenario eavesdropping channel state information is usually unknown. Therefore， this article considered a more realistic situation and aimed at the scenario of unknown eavesdropping channel state information， considering the existence of a legitimate receiver and multiple illegal eavesdroppers. From the perspective of power allocation， combined with beamforming， a multiple RISs auxiliary artificial noise and power allocation scheme was designed. Through the alternate optimization of the base station beamforming matrix and the RIS reflection coefficient and dichotomy， the power required to send confidential information to the legitimate user at the base station could be reduced， under the premise that the signal quality of the legally received user remains unchanged. Then signal to interference plus noise ratio of the received signal was reduced， by which the security capacity of the system was increased. Meanwhile， considering that the cooperative assisted confidential communication of multi-RIS may not be necessarily optimal， and may increase the complexity of system optimization， a multi-RIS selection scheme was proposed to improve energy efficiency while maintaining the secrecy of the communication system. The complexity of the designed algorithm was analyzed to prove the feasibility and effectiveness of the designed scheme. Then， the artificial noise scheme assisted by the joint beamforming RIS was simulated and analyzed. Finally， it is concluded that when the eavesdropping channel state information is unknown， rational joint optimization of RIS and base station beamforming for power allocation， combined with multi-RIS selection technology， can suppress the signal quality at the eavesdroppers， improve the energy efficiency of the communication system， and finally achieve confidentiality communication.