Sand tank experiments were performed along with on-site supplier experiments in order to obtain design factors for the raw-water supply system in floodplain filtration. Results of the sand tank experiment elucidated that the infiltration rate was approximately proportional to the soil permeability and was not significantly influenced by short periods of rest. The average daily infiltration rate calculated by taking both flood and rest periods into account increased with increasing flood period, and was observed to reach an asymptote. Under the conditions of this study, the maximum infiltration rates obtained for both Daegu and Mulgeum soils with 15 min/ 30 min of rest/flood periods were 6.3 m/day and 1.4 m/day respectively, which were 42% and 70% of their hydraulic conductivities, respectively. The process of soil filtration resulted in a gradual decrease of hydraulic conductivity; a decrease of 27% was observed for the soil of Mulgeum over a period of 8 days. From the data obtained from the supplier experiment, it was evident that the radius of the flooded area increased as the supply rate increased for soils of Gumi and Sangju, however, there was an inverse correlation between hydraulic conductivity and the rate of increase in the radius. Results also showed that the time required to cover the entire soil surface with water, in other words, the time to reach the maximum flood radius from the commencement of the water emission was as short as 3 to 4 minutes for all the soils. Also, the average infiltration rate for the entire flood period did not change significantly when the rest period was shorter than an hour.