Congestion is one of the most well-known problems in data networks. The congestion occurres when the number of packets being transmitted through the network exceeds the packet handling capacity of the network. The aim of AQM is to keep the queue size at a certain desired level at which network performance falls off dramatically. This situation leads to long time delays, and may even result in instability in the network. The purpose of this paper is to design controller for congestion control of the TCP network using Finite Spectrum Assignment (FSA) technique and comparison with other conventional methods including Proportional-Integrator (PI), Random Early Detection (REM) and H∞. To evaluate the performance of the proposed controller, different experiments are conducted through NS-2 simulations by considering variations in network dynamics.