The flow duration curve (FDC) serves as an essential tool for analyzing streamflow variability and supporting effective river management. However, constructing FDCs in ungauged basins presents a significant challenge due to the lack of sufficient data. This study leverages data-driven approach to predict FDCs in ungauged basins, thus offering practical solutions for improving hydrological forecasting and enhancing water resource management. The research aims to identify the key hydrologic, meteorological, and topographic factors influencing FDCs, and by evaluating different combinations of predictor variables, it assesses the influence of various precipitation metrics on flow predictions while comparing the performance of data-driven models. The study predicted low (Q80%, Q90%, Q95%), medium (Q30%, Q40%, Q50%, Q60%, Q70%), and high flows (Q5%, Q10%, Q20%), including extreme low flows (Q95%) and extreme high flows (Q5%). Feature importance analysis highlighted the watershed area and precipitation as critical for high flow predictions, and land use and basin characteristics influenced medium and low flows. Scenario testing confirmed that including all variables resulted in the most accurate predictions. Interestingly, variations in precipitation metrics had minimal impact on model performance, suggesting the prominence of other predictors. These results emphasize the potential of data-driven approaches in improving FDC predictions, particularly in diverse hydrological contexts where conventional methods fall short. This study highlights the potential of advanced hydroinformatics techniques to predict FDCs in ungauged basins, improving the accuracy of hydrological forecasting and water resource management through innovative, data-driven methodologies.Funding: This work was supported by Korea Environment Industry & Technology Institute(KEITI) through the Water Management Project for Drought, funded by Korea Ministry of Environment(MOE) (2022003610004).