Since slurry phase bioremediation is a promising treatment for recalcitrant compounds such as 2.4,6-trinitrotoluene (TNT), a statistical study was conducted for the first time to optimize TNT removal (TR) in slurry phase. Fractional factorial design method, 2(IV)(7-3), was firstly adopted and four out of the seven examined factors were screened as effective. Subsequently, central composite design and response surface methodology were employed to model and optimize TR within 15 days. A quadratic model (R-2 = 0.9415) was obtained, by which the optimal values of 6.25 g/L glucose, 4.92 g/L Tween80, 20.23% (w/v) slurry concentration and 5.75% (v/v) inoculum size were estimated. Validation experiments at optimal factor levels resulted in 95.2% TR, showing a good agreement with model prediction of 96.1%. Additionally, the effect of aeration rate (0-4 vvm) on TR was investigated in a 1-liter bioreactor. Maximum TR of 95% was achieved at 3 vvm within 9 days, while reaching the same removal level in flasks needed 15 days. This reveals that improved oxygen supply in bioreactor significantly reduces bioremediation time in comparison with shake flasks. (C) 2011 Elsevier B.V. All rights reserved.