In the battlefield communication confrontation, the reasonable distribution of interference parameters has always been a challenging task. This article is allocated based on strengthening the interference power, interference waveform and interference target of the interference party. Under the premise of ensuring the interference effect, it saves resource consumption and improves the utilization rate of resources. Specifically, the interference parameter distribution problem is constructed as a multi-smart task with complete collaboration, and the SA-QMIX algorithm with centralized training and distributed decisions can be used Introduce the maximum entropy method and MHA to make the smart parts more effective in some observation environments. The simulation results show that the SA-QMIX algorithm used in this article is compared to the traditional QMIX algorithm when the interference parameter distribution can increase the interference success rate by 5% by reducing the interference power of 1.5 dBm, and the convergence of the algorithm is faster.