Abstract: Network technology is very popular now because most offices and households use the internet for communication and carry out their day-to-day transactions. With the present high demand for internet users, they may experience congestion because of the increase in the number of users and demand on the network. Hence, controlling congestion for packet-switched data on the wired and wireless network using an artificial intelligence technique is very important to replace the current TCP protocols. This research used the neural network technique to realize this goal by training the model with some data set values which its result was later compared with the predicted result. This research works demonstrate great potential to control congestion for packet-switched on both wired and wireless networks. MATLAB was used for both the training of genetic programming and the ANN simulation for the prediction. When the message sent is higher than the capacity of the router, it can cause congestion. Therefore, this application is designed to be installed on the server to control congestion before its occurrence and congestion prediction. If some of the packets are queuing up, the application controls it by passing a message to the sender that congestion is about to occur along with the link. Therefore, the sender adjusts the speed of transmitting packets by reducing the level of the message the sender is sending and the flow of packets is then regulated.