Abstract: Recently, adaptive lattice filtering is finding increasing use in a variety of signal processing applications. Lattice structures offer a number of potential advantages over transversal filters. In this paper starting from the polynomial approximation of the Wiener filtering, we introduce the Szego polynomial. Then, it is shown that lattice structures could be derived naturally from transversal filters. Furthermore, the principal properties of lattice structures are discussed. Finally, two adaptive lattice algorithms are briefly presented.