People engage in visual search, finding a target among nontargets, all the time. Because of the limited attention capacity, only a few items can be processed. The nontargets compete for processing with the target; the more similar items, the stronger competition. Some human behavioral and monkey neural studies have shown that selective attention operates differently when the target and nontarget (T-N) competition are different. However, it is still unclear and controversial how neural mechanisms of selective attention operate in high and low competitive searches in human neural studies. I addressed this question by manipulating T-N similarity, assessing two neural correlates of selective attention using electroencephalography (EEG) and magnetoencephalography (MEG): (1) the N2pc effect, reflecting target selection, (2) the alpha lateralized effect, reflecting target prioritization and nontarget suppression, and exploring the effective connectivity of brain areas. In time-course information, for high similarity, the alpha lateralized effect occurred first, followed by the late N2pc effect (280 ms), and finally followed by the alpha lateralized effect, caused by decreasing power on the contralateral side, reflecting target prioritization. For low similarity, the early N2pc effect occurred (200 ms), followed by the alpha lateralized effect, caused by increasing power on the ipsilateral side, reflecting nontarget suppression. In the spatial and the effective information, I found the N2pc effect involved both top-down and bottom-up interaction for high similarity but only a bottom-up bias for the low similarity among four brain areas. Furthermore, the alpha lateralized effect, reflecting target prioritization, originated from the intraparietal sulcus (IPS). The effect, reflecting target prioritization, originated from the frontal eye field (FEF) and the temporal lobe. In summary, this dissertation provides spatiotemporal neural evidence to show that selective attention induces in different ways, depending on the competitive level.