In this paper, ten-year historical data of water levels recorded at Sto. Niño, Marikina station of MMDA-EFCOS were analysed and processed to determine the number of times per year (annual frequency) that critical levels of the Marikina River near the Sto. Niño station were reached and for how long (duration). Probability distributions for the annual frequency and duration were then fitted to the samples obtained. Monte Carlo simulation was applied in order to generate possible realizations of the random variables. Summary statistics were then obtained from the simulated values. Finally, backtesting using historical data of water levels after the period of model development was performed to check the validity of the models. The results showed that the models obtained were reliable. The results of this study may be used to guide the local government of Marikina in planning the needed resources in order to sufficiently respond in times of flooding incidents.