Modern networks employ congestion and routing management algorithms that can perform routing when network routes become congested. However, these algorithms may not be suitable for modern military Mobile Ad-hoc Networks (MANETs), more specifically, airborne networks, where topologies are highly dynamic and strict Quality of Service (QoS) requirements are required for mission success. These highly dynamic networks require higher level network controllers that can adapt quickly to network changes with limited interruptions and require small amounts of network bandwidth to perform routing. This thesis advocates the use of Kalman filters to predict network congestion in airborne networks. Intelligent agents can make use of Kalman filter predictions to make informed decisions to manage communication in airborne networks. The network controller designed and implement in this thesis will take in the current and predicted queue size values to make intelligent network optimization decisions. These decisions will enhance the overall network throughput by reducing the number of dropped packets when compared with current static network and MANET protocols.