Abstract: In order to ensure the power supply stability under large-scale new energy access, the power grid needs to perform real-time and accurate monitoring of distributed wind and solar power stations. However, the terminal access control in distributed wind and solar power stations currently faces several challenges such as information uncertainty in complex environment, relay selection conflict, and channel co-frequency interference. In this paper, a low-latency intelligent terminal access control algorithm based on co-frequency interference-aware and empirical inverse selection improved UCB is proposed. The empirical inverse selection mechanism is used by relays to solve the relay selection conflict problem, and the channel selection strategy is dynamically adjusted by terminal s through the channel co-frequency interference awareness based on current feedback reward. The case study verifies the proposed algorithm can reduce the transmission delay by 38.57% and 57.43% compared with two existing algorithms.