I have proposed a novel dip-separated structural filtering approach that cascades a new empirical mode decomposition based adaptive dip filter and one structural filtering approach. The seislet thresholding is used as a simple example for structural filtering. The principle idea of the proposed approach is to adaptively separate complicated seismic profiles that contain dip conflicts into separated profiles that contain no dip conflicts. The dip separation is achieved by the empirical mode decomposition based dip filter. Because of the dip decomposition, plane-wave destruction can obtain more precise slope estimation on each dip component and can help to obtain better compression performance for the seislet transform than the traditional implementation. Synthetic and field data examples show that the proposed approachcan help remove more random noise while preserving the energy of useful signals.