The accumulation phenomenon has been one of the major problems in computer networks in recent years, which is evident in the need to control this phenomenon in view of the growing Internet and its growing pace.Traffic on computer networks occurs when the amount of data input to a node in a network is greater than the output of that node. So far, there are many ways to control or prevent the phenomena of accumulation in computer networks that are mostly intuitive.These methods are examined from two different perspectives: in the first method, controllers are in the transmitter and the rest of the network is considered as a system in this approach. In the second method, the controller is in the router and the rest of the network is considered as a system. In this study, we present a model for predicting communications that occur in the next steps, and in this way, parts of the network with a high potential for congestion are already identified, and thus we can have better management. Consequently, by preventing, we will increase the loss of data due to network overcrowding.