Eleven soil samples (contaminated and non-contaminated top soils and aquifers) from seven different locations in Belgium were examined in lab-scale batch microcosms simulating in situ conditions for their indigenous capacity to biodegrade methyl tert-butyl ether (MTBE). The effect of implementing nutrients or additional oxygen and of the presence of co-contaminants on MTBE degradation was investigated. All soils showed rapid degradation of benzene. On the other hand, only one site, historically contaminated with oxygenated fuel, provided soil samples showing relatively fast MTBE biodegradation. These soil samples originated from four different depths from the vadose and saturated zone. MTBE degradation kinetics differed between the samples of the saturated and non-saturated zone and depended on the implemented conditions. MTBE-biodegradation in the samples from the non-saturated zone started after a very short lag-phase (<7 days), while long lag-phases (up to 270 days) were obtained with the other samples. Addition of extra nutrients stimulated MTBE degradation kinetics in microcosms containing the saturated soil samples while inhibiting effects were seen in the case of non-saturated soil samples. In contrast, implementing dissolved oxygen concentrations of 9.5 and 11.5 mg l−1 led to lower degradation kinetics compared to 8 mg l−1 in microcosms containing saturated soil samples, while stimulating effects were seen with the non-saturated soil samples. Addition of an extra carbon source like benzene or propane did increase in general the MTBE first order degradation rate constant. Differences in the eubacterial community composition between these depth samples were confirmed based on denaturing gradient gel electrophoresis (DGGE) patterns of PCR-amplified 16S rRNA gene fragments. The results of the presented study indicate that an aerobic MTBE biodegradation potential is not omnipresent in Belgian sub-soils.