Abstract: The de-noise processing of seismic data directly influence on the quality of subsequent data processing and on the accuracy of data interpretation. The former de-noise methods often affect the original signals while improve S/N ratio and so cannot keep the original appearance of the effective signals. The principle of LIFT is to get the signal model trace and eliminate it from the original data so as to separate noise from the data. Then, the separated noise data is de-noised again to get the residual data (the secondary signal data). Finally, the signal model traces and the residual data are reconstructed, showing that the LIFT method can not only improve the S/N ratio but also preserve the effective signal of the original data to a great extent. The application of the LIFT method in Daqing oilfield shows that the signal can be well preserved and the noise which cant be seen in the data by visual examination is clearly showed in the separated noise record. Comparison of the profiles from the LIFT method with that from the conventional de-noise method indicates that the LIFT can well suppress the noise, increase the S/N ratio, and enhance the continuity of the events. As a result, the energy distribution of profile in lateral becomes more reasonable.