In wireless communication systems, a signal that travels from transmitter to receiver gets corrupted due to various interference effects, distortions, noises and multipath problems. Signal also undergoes scattering for each major path. Multipath propagation causes fluctuations in amplitude, phase and angle of arrival of the received signal giving rise to fading of the signal. Conventional receivers are unable to compensate for nonlinearities and other complex problems. Some computational intelligence is required to deal with fading multipath issues. Intelligent systems are able to identify the problem and take necessary corrective action in the light of ongoing events of their own. Neural networks are non-rule based intelligent systems with adaptive feature that aim to perceive and comprehend significance of data with which they are trained. This thesis presents a neural network based; high power amplifier (HPA) characteristics independent, adaptive method for compensation of nonlinear distortions for OFDM based wireless communication system.