In order to provide a fully automatic antenna adaptive adjustment scheme with better performance in all aspects, and achieve better coverage effect while using lower maintenance cost, the key technologies of intelligent adjustment system design of adaptive antenna feed system of 5g-based station are studied from the perspective of signal radiation direction adjustment of 5g antenna panel. An adaptive adjustment strategy for base-station antenna based on deep reinforcement learning is proposed. The 5g base stations adaptive days feed system is designed, which is based on deep reinforcement learning techniques, using telecom RSRP coverage map as a data source, and it can obtain the current state of the observed values and automatically analyze data and adjust the antenna panels. In a virtual environment, the system based on reinforcement learning is simulated and trained, and the results are in line with expectations.