In Taiwan, a substantial amount of thin-film transistor liquid crystal display (TFT-LCD) wastewater is produced daily due to an increasing production of the opto-electronic industry in recent years. The main components of TFT-LCD wastewater include dimethyl sulphoxide (DMSO), monoethanolamine (MEA), and tetra-methyl ammonium hydroxide (TMAH), which are recognized as non-or slow-biodegradable organic compounds and limited information is available regarding their biological treatablility. This study was conducted to evaluate the long-term performance of two bioreactors, anaerobic-aerobic (A/O) sequencing batch reactor (SBR) and aerobic membrane bioreactor (MBR), treating synthetic TFT-LCD wastewater containing DMSO, MEA, and TMAH with different loadings. For the A/O SBR, the influent wastewater was composed of 800mg MEA/L, 430 mg DMSO/L, and 90 mg TMAH/L, respectively. After reaching steady-state, SBR was able to achieve more than 99% degradation efficiencies for the three compounds examined. For the case of aerobic MBR, the influent wastewater was composed of 550 mg MEA/L, 270 mg DMSO/L, and 330 mg TMAH/L, respectively, and degradation efficiencies for the three compounds achieved more than 99%. Although both different reactors shared similar and satisfactory degradation efficiencies for DMSO, MEA, and TMAH, the microbial ecology of these two reactors, as elucidated with molecular methods, was apparently different. The 16S rDNA-based cloning/sequencing results indicated that the dominant sequences retrieved from the aerobic MBR, including Hyphomicrobium denitrificans, Hyphomicrobium zavarzinii, Rhodobacter sp., and Methyloversatilis universalis, showed a clear linkage to their physiological properties of DMSO and TMAH degradation. On the other hand, Zoogloea sp., Chlorobium chlorochromatii, Agricultural soil bacterium, and Flavosolibacter ginsengiterrae were proliferated in the A/O SBR Run1, while Thiobacillus sp., Nitrosomonas sp., Thauera aromatica and Azoarcus sp. became dominant in Run2. Furthermore, the sequences retrieved from different reactors were used to establish the terminal restriction fragment length polymorphism (TRFLP) fingerprint methodology for monitoring the dynamics of dominant degrading bacteria in the aerobic MBR treating TFT-LCD wastewater.