The aim of this study was to evaluate how routinely performed phosphorus release tests could be used when modelling enhanced biological phosphorus removal (EBPR) using activated sludge models such as ASM2d. A pilot plant with an extensive analysis programme was used as basis for the simulations. Without any calibration the prediction of phosphorus removal was poor and the initial release rates from the simulations were not similar to those found from the laboratory tests. A period with low organic loading was chosen as a calibration period. In this period averages of daily influent measurements were used as influent parameters. First, calibration was performed in order to fit effluent COD and MLVSS in the sludge. Next, the phosphorus content in the sludge was decreased to the measured level by decreasing the fermentation rate. Finally, the initial phosphorus release rate was calculated from a simulated batch test and the PHA uptake rate was increased to fit this release rate with the average initial rates from laboratory batch tests performed during the period. The calibrated model was verified with data from the subsequent period where acetate was dosed.