Abstract: For prestack seismic data with low signal-to-noise ratio(SNR), the effect of traditional AVO rendezvous analysis and intercept gradient product attribute(P×G attribute) may not be ideal. In order to extract the information of AVO attribute effectively, this paper starts from the prestack partial superposition body, according to the relative relation of the far and near track part superposition body, the far and near attribute factor similar to the P×G attribute was constructed, in which the near track stack reflects the intercept attribute, and the near-far stack difference reflects the intercept attribute, which is similar to the P×G attribute. And it is verified by theoretical model that it can replace P×G attribute in AVO analysis of actual seismic data. At the same time, the signal-to-noise ratio of partial stack data is significantly higher than that of single-channel set. Finally, taking a gas reservoir in deep water area in Qiongdongnan basin as an example, it is shown that the difference attribute of far and near track attribute can effectively solve the full star phenomenon of actual gathers, thus providing a new way to study the AVO attribute analysis of seismic data with relatively low SNR.