Abstract: In wireless networks, when using traditional game studying physical layer security, the energy-limited transmission nodes tend to choose a non-cooperative strategy in order to maximize their own secrecy rate, resulting in reduced network secrecy rate. To solve this problem, this paper presents a method for physical layer security cooperation based on evolutionary game. Firstly, this study defines strategies (sending artificial noise or signal) and benefits (secrecy rate under different strategy combination) according to evolutionary game; then, the transmission nodes adjust strategy to maximize benefits based on current network state and difference between cooperation utility and average expected utility; finally, the conditions that the transmitting nodes can achieve stable cooperation, are obtained, and the network evolution from an unstable state to a stable collaboration state improves the secrecy rate of the system. The simulation and analysis results show that under the Gaussian channel conditions, compared to traditional game method, the proposed methods network secrecy rate can be improved 1 bit/(s Hz).