To determine if anaerobic microniches are created in granular sludge exposed to oxygen where the refractory organics reduction in coking wastewater could occur and to assay the feasibility of the highly efficiently and stable operation of the micro-aerobic EGSB reactor for simultaneous removal of COD and NH 3 -N treating actual coking wastewater at 25-30℃, both anaerobic and microaerobic operations of the EGSB reactor were considered. And the granular characteristic of activity, settling, and size was also investigated in this paper. By controlling the oxygenation rate in the aeration column, different concentrations of dissolved O 2 were generated in the circulating fluid which supplied dissolved O 2 to the granule sludge bed in the EGSB reactor and to generate micro-aerobic environments. The results showed that, at 25-30℃, treating actual coking wastewater with 1280-2024mg⋅L -1 COD concentration, 50-112mg⋅L -1 NH 3 -N concentration, the anaerobic EGSB reactor could only have 40.1%, 56.7% average COD removal efficiencies and 8.3% average NH 3 -N removal efficiencies (the negative NH 3 -N removal was usually appeared and even low to -77.8%) with 1.3L. -1 , 0.6L -1 influent flow and 9.2h, 20h HRT, respectively. But the microaerobic EGSB reactor could have average COD and NH 3 -N removal efficiencies of 83.0% and 67.1%, respectively. Supplement dose of Oxygen for the microaerobic EGSB reactor was controlled by redox potential (ORP), and the ORP in the inlet of the micro-aerobic EGSB reactor should be kept at about +35mV. Oxygenation could shift the population distribution to smaller or larger particles compared to the anaerobic EGSB reactor and decreased the settled velocity of the granules (low to 12m.h -1 ), but without any sludge washout. And the micro-aerobic EGSB reactor could keep about 38.0gMLSS.L -1 sludge concentration and 0.65 MLVSS/MLSS. The methanogenic activity of the granular sludge in the micro-aerobic was not decreased, and even higher than in the strictly anaerobic EGSB reactor.