The use of urban stormwater quality models necessitates the estimation of their outputs’uncertainty level.The results of the application of a Monte Carlo Markov Chain method based on the Bayesian theory for thecalibration and uncertainty analysis of a storm water quality model commonly used in available softwareare presented in this paper. The tested model estimates the accumulation, erosion and transport of pollutants onsurfaces and in sewers using a hydrologic/hydrodynamic scheme. The model was calibrated for 4 different initialconditions of in-sewer deposits. Calibration results showed a large variability in the model’s outputs in function ofthe initial conditions and demonstrated that the tested model’s predictive capacity is very low