Sound-induced flash illusion is a typical auditory dominance phenomenon. This illusion can be divided as fission illusion (when a single brief visual flash is accompanied by two auditory beeps and participants report they perceived two flashes) and the fusion illusion (where a double flash is accompanied by a single auditory beep and participants report they perceived a single flash). Compared with fusion illusion, the effect size of fission illusion is larger and stronger. Prior studiesfound that endogenous attention couldinteract with multisensory integration. Furthermore, adirect evidence support that modal-based endogenous attention (selective attention vs. divided attention) influenced on sound-induced flash illusion, spatially for fission illusion. However, the relationship between endogenous spatial attention and sound-induced flash illusion is still uncertain. In addition, visual field is an important factor in visual processing, because there is a significant difference between upper and lower visual field in some tasks. Therefore, in the present study, we focused on the effect of endogenous spatial attention and visual field on sound-induced flash illusion. In experiment 1, we used a 2 (endogenous spatial attention: selective attention vs. divided attention) × 2(visual field: upper vs. lower) factorial design, and the factor of endogenous spatial attention was manipulated between blocks. The task for participants was to judge the number of flash that occurred in a certain place or any place. However, in experiment 1, even if the attention was manipulated by instruction, the flash would occur in upper or lower visual field randomly, so whether the uncertainty of the place that the flash occurredwas another influential factor in the paradigm? To exclude this possibility, we added an arrow, which could indicate the place that the flash would occur before the flash appearance in experiment 2. So experiment 2 was also a 2(arrow validity: valid vs. invalid) × 2(visual field: upper vs. lower) factorial design. The valid arrow directed upper or lower visual field in a trial, and invalid arrow directed upper and lower visual field at the same time. Moreover, all trials presented randomly. From the results of accuracy (ACC) in experiment 1, we found that fission illusion was more robust than fusion illusion, and endogenous spatial attention and visual field did not influence sound-induced flash illusion. However, they could contribute together to this illusion. Specifically, when attentiondivided, the illusion was larger in upper visual field than in lower visual field. Nevertheless, under selective attention condition, there was no difference between upper and lower field.In addition, the ACC of experiment 2 indicated the illusion influenced by visual field rather than endogenous spatial attention. Moreover, there was no significant interaction between endogenous spatial attention and visual field. In other words, the addition of the informative arrow contributed to the upper-lower asymmetry effect, and weakened the interaction between endogenous spatial attention and visual field. Taking the results of experiment 1 and 2 together, we suggest that the fission illusion is independent on endogenous spatial attention and visual field, but influenced by coefficient of the two factors when there is no arrow, which can indicate the place that flash occurs. However, the arrow can increase the upper visual field advantage in fission illusion, and decrease the interaction between endogenous spatial attention and visual field.