Changing climate, human interventions to natural water flow pattern, haphazard urbanization etc., are the reasons for intense flood even after development of so many structural measures of overflow control. Kulfo River basin is situated in relatively dry southern area of the Ethiopia and is still under geographical modification with hilly topography and impervious soil texture. The concern of the present research is to simulate flood episode in order to develop flood management strategies to reduce disaster. The complexicity of natural hydrological phenomenon and dependent random variables can be better expressed considering it as stochastic process. Flood (maximum river flow) forecasting on the Kulfo River with monthly runoff data using stochastic ARIMA, Time Series model was developed for warning purposes. The analysis of seasonally varying time series of discharge data has revealed that a higher order ARIMA model may produce excellent results for three to six months forecast.