Tinnitus is one of the most common symptoms affecting people all over the world. In the absence of an established cure many individuals are not only faced with the need to adjust to the sensation of the tinnitus noise, but also with psychological comorbidities. In recent years, different studies have been directed to elucidate the psychophysiological mechanisms that are involved in the tinnitus decompensation. From these, special emphasis has been placed on studies related to attention and habituation, which accordingly play a crucial role in current tinnitus therapy approaches. In spite of such progress, the relationship between selective attention and the tinnitus decompensation with respect to large-scale neural correlates is still not well understood. In order to address this issue, we propose an integrative multiscale modeling approach for studying neural correlates of auditory selective attention in the tinnitus decompensation. Computational simulations based on our model confirmed electroencephalographic human data of both auditory selective attention and the tinnitus decompensation. It is concluded that the proposed methodology represents a promising approach to give insight into the neurodynamics of auditory selective attention in the tinnitus decompensation.