Facing the goal of "peak emission and carbon neutrality power distribution carbon monitoring wireless sensor network distinguishes itself with unique advantages of flexible deployment and strong invulnerability, and can provide communication support for carbon monitoring service, the integration of grid and carbon market, and achieving the peak emission and carbon neutrality. To satisfy the low-energy consumption and ultra-reliable transmission requirements, based on the principle of reinforcement learning, this paper proposes a power-distribution-sensing technology promoting strategy oriented for peak emission and carbon neutrality, which can minimize the network energy consumption and packet loss by jointly optimizing relay selection and power control. Compared with the existing Maximum Channel Gain of Data Links (MaDs) method, simulation results show that the proposed strategy can reduce the transmission energy consumption and the packet-loss ratio by 26.47% and 7.19%, respectively. The proposed work contributes to the low-energy consumption and ultra-reliable data transmission and thereby laying a communication foundation to realize peak emission and carbon neutrality.