Abstarct: Seismic attributes are widely used as a tool for better interpretation of seismic data with their implicit information. One of the factors affecting the performance efficiency of seismic attributes is the quality of the input data, so that the results obtained from applying an identical attribute on a single seismic datum which has been processed with different methods may differ. In this research, two methods – conventional stacking and common reflection surface stacking – were used to obtain zero offset section. Conventional stacking depends on velocity model, and due to the selection of one line for stacking, it does not have a high signal to noise ratio and continuity of events is low in it. This is while in common reflection surface stacking, due to the selection of one surface for stacking, signal to noise ratio in the final section is more and events have higher continuity and quality. To calculate the attributes, it is better that the input data are migrated. In this research, sections obtained from conventional stacking and common reflection surface stacking were migrated by Kirchhoff method, then the intended seismic attributes were applied on the migrated sections. Seismic attributes used in this paper are chaos, coherency and texture attributes. Chaos and coherency attributes are very efficient in identifying faults and discontinuities; texture attributes are used in separating areas with different textures such as salt dome. Finally, seismic attribute sections obtained from the two stacking methods were compared. Results showed that due to the selection of stacking surface, the common reflection surface stacking method acts better than the conventional stacking method, and the continuity and quality of sections obtained from common reflection surface stacking method is more than the conventional stacking method, which leads to more efficiency in studying and interpreting seismic attributes, compared to conventional stacking sections.