We present the current state of veto studies in TAMA300 by monitor signals of the interferometer and its environment. In GW data analysis, fake GW events may bury real GW events or worse upper limits on the event rate. Thus, fake-event rejection is an important issue. In general, we can reject these fake events by the monitor signals, since these fake events are induced due to detector instabilities. However, using all monitor signals for the fake-event rejection would increase the accidental rejection probability and dead time without improving veto efficiency, since all monitor signals do not have correlations to the detector instabilities. Here, we analyze coincidences between the main and selected monitor signals with the optimal parameters for the fake-event rejection. Then, coincident events are rejected as the fake events. For the signal selection and parameter optimization, we systematically investigate the correlations with the detector instabilities. As a result, we achieved 30–99% veto efficiency using ten selected monitor signals with the 3.2% accidental rejection probability and 0.2% dead time.