Abstract: The current work confirms that the derived moment approach is a suitable replacement for the curve fitting method when runoff data are not readily available. Specifically, annual runoff frequency was estimated using climate variables associated with hydrological processes. The difficulties involved in assessing the annual runoff frequency in the ungauged basins and the solutions to these problems are discussed. Here, 20 basins in Jiangxi Province were examined using the two annual rainfall—runoff models established based on the Schreiber equation and Fu-Zhang equation. Three relevant conclusions were drawn. First, the frequency distributions of annual runoff estimated using the derived moment approach fit the empirically observed frequency of the annual runoff perfectly. Second, the Fu-Zhang model was found to perform better than the Schreiber model when the basins' interannual changes in soil moisture content were not known. Finally, the parameter ω in the Fu-Zhang model was estimated using the inverse distance weight interpolation method through the calibrated value ω of neighboring basins whose runoff data were available.