Abstarct: Recently, Transmission Control Protocol (TCP) is the most widely used and most popular transport protocol. The TCP with reliable data distribution and congestion control constitute the basis of current internet over Internet Protocol (IP). When the total amount of data entering the network exceeds of the network capacity, congestion occurs. During the congestion in network the router queues get full, so packets passing of the routers suffer from considerable delays. If congestion still persists, the length of queue may be greater than buffer size, then buffer overflow. Representing mathematical model of TCP network, different control methods for active queue management (AQM) are proposed. One way to deal with the delay is use of Smith predictor. But major disadvantage of Smith predictor which limits its application is the problem of its sensitivity to modeling errors. According to the obtained mathematical model for the network TCP, mismatch between the model and the process is inevitable. Hence, the purpose of this thesis is active queue management for TCP network using H∞ control theory baxsed on Smith Predictor. The results of the proposed method are studied and compared with the classic controller and another method (FSA) for TCP network congestion control. The most important feature of the proposed method is that the designed controller is robust against network dynamics including changes in type of network traffic and traffic load.