Abstract: This paper studies the influencing factors of seismic data on prestack imaging quality and suggests that the prestack data regularization should be performed by both homogenization processing of data spatial sampling and data energy. By means of borrowing nearest traces, the normalization of data spatial sampling can improve the signal to noise ratio of migration imaging; using triangulation method for normalization of data amplitude energy may decrease the migration noise and strengthen AVO attribute. By joint regularization processing of prestack seismic data and application to Nanpu area, the imaging quality after regularization is much more improved than that before it, including improvement of signal to noise ratio in shallow and suppression of noise in deep zone as well as amplitude preservation