<p>Hydrological modeling in ungauged basins, often characterized as data-sparse regions, presents various challenges. Additionally, the wide range of rainfall-runoff models with different levels of parameter complexity often complicates model selection. This study evaluates five widely used rainfall-runoff models in Indonesia: NRECA, FJ-Mock, HBV96, NAM, and Sacramento. These models are categorized into three groups based on the number of parameters: small number (NRECA and FJ-Mock), moderate number (HBV96 and NAM), and large number (Sacramento). The Waluh Catchment in Central Java, Indonesia, is used as a case study, where the available data include only monthly precipitation, monthly discharge, and potential evapotranspiration (PET) from 2003 to 2009. Model calibration and validation were conducted manually using data from 2003–2006 and 2007–2009, respectively. Model performance was evaluated using four objective functions: Nash–Sutcliffe Efficiency (NSE), Linear Correlation Coefficient (R²), Relative Volume Error (RVE), and Kling–Gupta Efficiency (KGE). All models produced similar simulated discharges that closely matched the observed data under low flow conditions, with minimum R², NSE, RVE, and KGE values of 0.85, 0.59, 0.13, and 0.59, respectively. Flow Duration Curves (FDCs) analysis also confirmed minimal differences between observed and simulated discharge. Notably, models with fewer parameters required less calibration time, whereas models with more parameters demanded extensive input data to justify the calibrated values. These results suggest that models with fewer parameters can produce outcomes comparable to more complex ones. Therefore, in data-sparse catchments, simpler models are more practical and reliable for rainfall-runoff simulation and water resource planning.</p>