Although the volterra filter is known to be an efficient adaptive filter in a large variety application for the non-linear system, but its computational complexity is also very high, especially as the orders of the filter increase. To overcome the computational complexity of the Volterra filter, a novel adaptive filter using layered bilinear architecture is proposed in this paper. Compared with the conventional second-order Volterra filter and Direct Bilinear adaptive filter, the layered bilinear adaptive filter exhibits a slightly better convergence performance in terms of convergence speed and steady-state error.