Abstract: ‍ ‍The channel openness of wireless communication systems made them extremely vulnerable to external malicious interference， which was difficult to guarantee the quality of communication links. This paper designs a multi-domain joint interference avoidance decision-making method based on deep reinforcement learning. This method combined the anti-interference means of the frequency domain， power domain， and modulation coding domain to avoid interference， and realized reliable communication while considering system performance at the same time. Firstly， the joint intelligent interference avoidance problem was modeled as a Markov Decision Process （MDP）， the action space included switching channels， adapting power， and changing the modulation and coding method. Then， the proximal policy optimization algorithm based on the objective clip （PPO-Clip） was used to solve the optimal joint interference avoidance strategy of the system. The PPO-Clip algorithm was iteratively updated with a small number of samples in multi-turn training， which avoided the problem that the step length is difficult to determine and the difference in policy updates is too large in the policy gradient algorithm. Finally， the effectiveness and reliability of the proposed algorithm are verified under the environment of sweep interference， random sweep interference， and intelligent blocking interference， respectively.