As distinct from the USLE/RULSE for which only rainfall rate is required to quantify the climatic influence on soil erosion, information on runoff rate in addition to rainfall rate is needed to drive all the process-based soil erosion models. For numerous runoff and soil loss plots around the world, only event-based runoff amounts have been measured regularly. To estimate soil erodibility parameters for these process-based soil erosion models and to facilitate their use, methods for estimating the runoff rate for given runoff amount are clearly needed. Three one-parameter infiltration models were examined, and estimated and observed runoff rates were compared for six bare plots in the tropical and subtropical regions of Southeast Asia and Australia. The model which characterizes the spatial variability in the maximum rate of infiltration performed best in comparison with other models assuming either a constant or a proportional infiltration rate. The spatially variable infiltration model has little bias in estimated runoff rate and works equally well for data collected at different time intervals. At a typical time interval of 6 min, the relative bias of the estimated peak runoff rate is 6% and the mean absolute error is 4.6 mm/h. The ratio of root mean square error to peak runoff rate was found to decrease as the runoff coefficient increases.