Abstract: Resolution of seismic data is an important factor of constraining seismic prospecting precision, while high?resolution seismic data processing aims to properly recover high and low frequency spectrum information from seismic records and effectively broaden the bandwidth or frequency spectrum. The conventional approaches to get them are applied as follows: Deconvolution approach is based on seismic convolution model, which requires various assumptions of seismic wavelet, reflection coefficient, stratigraphic occurrence and excitation?receiving pattern; inverse Q filtering approach supported by mechanism of seismic attenuation can compensate and correct amplitude attenuation and phase distortion caused by earth filtering, which depends on the precision of Q estimation and the match degree between model and actual data; frequency recovery approach based on time?frequency spectrum allows to rational estimation of unsteady state wavelet’s amplitude and phase. All in all, these approaches aim to broaden the bandwidth in nature, thus resulting in two changes in seismic profile during high?resolution processing: one is most events become thinner and incremental (due to shorter wavelength); the other is some events become weaker or even disappear (due to smaller side lobes of wavelet). Compared to high frequency component, low frequency component has greater effect on enhancing seismic profile’s layering and improving inverse accuracy, and its recovery is much more difficult. Hence the low frequency component information should be paid more attention to in protection and recovery during such a processing in the future