The paper presents the adjustments carried out on the structure and in some default values of the kinetic coefficients of the ASM2d model in order to get an improved prediction for the experimental results obtained in pilot scale plants studies with different activated sludge treatment processes for carbon, nitrogen and phosphorous removal. In order to predict the high effluent filtered COD experimentally observed in high rate processes for carbon removal, a new model structure has been proposed, incorporating into the carbon model a soluble fraction of the slowly biodegradable substrate. Studies with the step feed and the alternating processes showed simultaneous nitrification-denitrification in aerated reactors which was predicted with increased values in the oxygen saturation coefficients for heterotrophic and autotrophic biomass. Both processes also showed loss of the denitrification capacity under unfavourable conditions, such us rains and low anoxic fraction, which required a very large decrease in the value of the reduction factor for denitrification so as to improve the predictions for effluent nitrate experimental results. Regarding phosphorus removal, the ASM2d model showed a satisfactory predictive capacity. For improved predictions of phosphorus release in anaerobic conditions, high values of the rate constant for storage of X(PHA) and low values of the anaerobic hydrolysis reduction factor were used. For phosphorus uptake in aerobic and anoxic conditions satisfactory predictions were obtained using the default values.