Hydrocarbons exploration is geared towards identifying complex geologic and stratigraphic features in different scales. Due to the potential of channel facies in hydrocarbon storage, identifying and depicting them has been a great challenge in stratigraphic interpretation. In recent years, seismic data interpretations cause difficulties for the interpreters due to the increase of the seismic data volume and the variety of seismic attributes. Integrating seismic attributes with different computational algorithms based on artificial neural network, provides more details of seismic events. In this study, a semi-automatic method based on seismic attributes integration has been used to map buried channels. Following seismic data preconditioning, the attribute set and the pick sets selected by interpreter were presented to the supervised neural network. The results showed that an improved image of the channels with relatively a high resolution is illustrated.