Abstract: In order to study the propagation effect of the precipitation uncertainties on the flood forecasting,the paper presents a methodology combining the fuzzy set theory with the genetic algorithm to quantify the effect of the uncertainty on the discharge simulation due to the precipitation magnitude and temporal-spatial distribution.The uncertainty of precipitation magnitude is represented by the fuzzy set concepts.The random disaggregation of precipitation into the shorter time step takes into account the uncertainty associated with temporal distribution of precipitation.The spatial variations of rainfall fields are represented by the different temporal disaggregation patterns within the sub-catchments.All the kinds of uncertainties are propagated through the fuzzy extension principle.Based on the methodology above,the TOPMODEL is used to simulate the flood events in the upstream area of the Xinning hydrological station.The results show that the uncertainty in the magnitude is more significant than the uncertainty resulting from the spatial-temporal distribution of precipitation.Moreover,there is no more difference between the precipitation uncertainty with time step of 1 hour and that with time step of 0.5 hour.It means the simulated time step of.hour is able to represent the temporal variation of precipitation in the study area.