Aerobic digestion is a desirable method of sludge stabilization used in, especially small, wastewater treatment plants. In this processendogenous phase will be occured, in which the death of organisms, cell lysis, and predation among the microorganisms cause the active biomass to constantly reduce. Mathematical modeling allows a large number of potential aerobic process designs to be tested and the performance of aerobic digesters to be evaluated. In applying such models, however, it is essential to use appropriate models that can neatly explain the kinetics of the process. In this thesis, the simple first-order model (Adams et al . 1974) and the Activated Sludge Models developed by IWA (ASM1 and ASM3) were evaluated. For this purpose, three batch reactors with initial TSS concentrations of 5,000, 10,000, and 20,000 mg/L and operating volumes of 10 L as well as three semi-continuous reactors with operating volumes of 5 L and mean solids retention times of 5, 10, and 20 days with corresponding daily feed rates of 1000, 500, and 250 mL/d, respectively, were operated to investigate their performance and to collect the experimental data required for model calibrations. The batch reactors and the semi-continuous ones were monitored for 70 and 60 days, respectively. Samples taken from the reactors were analyzed for their solids concentrations and their Chemical Oxygen Demand (COD). Experimental data revealed that the more concentrated the sludge, the lower its stabilization rate but the higher the ultimate efficiency of the digester. It was also shown that increased mean cell residence time in semi-continuous reactors led to higher reductions in biological activity, higher suspended solids degradation, and a more stabilized sludge produced. The measured data were then used to estimate the parameters of the models investigated and the best sets of parameters yielding the best fit of the predicted and measured data were identified for each model and for each reactor. For each of the batch and semi-continuous runs, representative sets of parameters were also identified whose total errors over all reactors were minimum. The parameter estimation results for batch and semi-continuous runs showed that higher sludge concentrations and mean cell residence times were associated with lower cell decay coefficient values. The error values for batch and semi-continuous runs calculated by ASM3 were found to be lower than those calculated by the simple first order model and by ASM1. Comparisons with experimental data proved ASM3 to be superior in its capability to predict concentrations in both batch and semi-continuous reactors.