Abstract: This study was conducted with column experiments, batch experiments, and bench-scale permeable reactive barrier (PRB) for monitoring the PRB in the relation between BTEX (benzene, toluene, ethylbenzene, and p-xylene) decomposition efficiency and the distribution of a microbial community. To obtain the greatest amount of dissolved oxygen from oxygen-releasing compounds (ORCs), 20-d column tests were conducted, the results of which showed that the highest average amount of dissolved oxygen (DO) of 5.08 mg l−1 (0.25 mg-O2 d−1 g−1-ORC) was achieved at a 40% level of CaO2. In the batch experiments, the highest concentrations of benzene and toluene in which these compounds could be completely degraded were assumed to be 80 mg l−1. Long-term monitoring for a PRB indicated that ORCs made with the oxygen-releasing rate of 0.25 mg-O2 d−1 g−1-ORC were applicable for use in the PRB because these ORCs have a long-term effect and adequately meet the oxygen demand of bacteria. The results from the DGGE of 16S rDNAs and real-time PCR of catechol 2,3-dioxygenase gene revealed the harmful effects of shock-loading on the microbial community and reduction in the removal efficiencies of BTEX. However, the efficiencies in the BTEX decomposition were improved and the microbial activities could be recovered thereafter as evidenced by the DGGE results.