Stochastic approach to the derivation of WWTP emission limits meeting probabilistically defined environmental quality standards (EQS) is presented. The stochastic model is based on the mixing equation with input data defined by probability density distributions and solved by Monte Carlo simulations. The approach was tested on a study catchment for total phosphorus (Ptot). The model assumes input variables independency which was proved for the dry-weather situation. Discharges and Ptot concentrations both in the study creek and WWTP effluent follow log-normal probability distribution. Variation coefficients of Ptot concentrations differ considerably along the stream (cv=0.415–0.884). The selected value of the variation coefficient (cv=0.420) affects the derived mean value (Cmean=0.13 mg/l) of the Ptot EQS (C90=0.2 mg/l). Even after supposed improvement of water quality upstream of the WWTP to the level of the Ptot EQS, the WWTP emission limits calculated would be lower than the values of the best available technology (BAT). Thus, minimum dilution ratios for the meaningful application of the combined approach to the derivation of Ptot emission limits for Czech streams are discussed.