In the information network centric warfare the network congestion problem becomes the bottleneck of warfare information transformation.In single neuron PID congestion control the neuron gain K is the main factor affecting the system performance of stability and overshoot which is totally depended on expertise and is highly subjective.In this paper the neuron gain K is adjusted dynamically combined with identification-free adaptive control algorithm.In addition in order to improve the quality of the controller the adjustment factor is introduced to adjust the learning rate of neuron weight.The simulation results in the tactical network verified the effectiveness and feasibility of the improved algorithm.