This paper focuses on Maximum Likelihood Sequence Estimate (MLSE), based on the widely study of basic principle, structure and algorithms of adaptive equalization.

As a kind of new optimized algorithm, neural network can substitute for landscape orientation filter in adaptive equalization, and it has been widely used in digital communication system.

Based on the analysis of wavelet packets and traditional adaptive equalization, an adaptive equalizer structure and its adaptive algorithm, based on wavelet packets transform, are presented.