An attempt was made to assess the use of the empirical washoff model in predicting the concentrations of total suspended solids contained in storm runoff in the 12 sub-catchments of agricultural, residential and forest land use in the Kranji Reservoir catchment, Singapore through (1) an application of the washoff independent from the buildup model, (2) considering initial mass on surface Bini as a calibration parameter, (3) considering washoff coefficient c3 as a variable rather than a constant, and (4) investigation of the sensitivity of washoff parameters c3, Bini and the washoff coefficient c4 with rainfall and runoff characteristics, land use and catchment size. The optimized model parameters c3, c4, and Bini were obtained using the Monte-Carlo simulation. The results showed that the washoff model can be applied to catchments of all three mentioned types of land use with reasonable fit. Bini for the agricultural catchments are significantly higher than that from the residential and forest catchments, which is in agreement with observations on the event mean concentrations of TSS. It was also found that the values of c3 and Bini for the same sampling station varied significantly with the rainfall depth (d). Bini was found to be correlated mainly with d, but also to a lesser extent with average intensity, peak runoff volume, and runoff volume. Significantly, Bini did not correlate well with ADP or with rainfall depth of the prior storm event. The results from this study show that the common modeling practice where the initial mass available for washoff Bini is simulated using antecedence dry period and previous rainfall depth should be reassessed, at least when applied to tropical catchments. The study also shows that washoff parameters depend on catchment size. The slopes of regression between c3 and Bini with d increase with catchment area, implying that washoff behavior in smaller catchments is less sensitive to changes in rainfall depth than larger catchments.