Minor component analysis (MCA) of lower dimensional data is related to many signal processing applications. MCA strives to extract the "minor" direction in the data space where the variance of the data is minimal, identify the way for dimension reduction and data compression. In this paper neural networks are used to estimate the minor component of signal. This component is used to determine the Direction of Arrival Estimation (DOA) of incident signals. These signals are considered to be emitted from their emission sources .The neural networks knowing “Hebbian-networks” are used to estimate the minor component directions from signal subspace. Narrow band signals are considered here and strike an array composed of M sensors. Simulation results are introduced to shown the performance of the adaptive neural networks to estimate signal components, a comparison of the results obtained from classical method and MCA method, is presented which shows the performance of MCA over classical methods, to estimate exact signal direction from noise subspace.