In this research, the hydrological response to changes in land use for the Quilmo River watershed (Southern Chile) is assessed. The assessment is performed using the Hydrological Simulation Program-Fortran code (HSPF). The hydrological response to predominant land uses of non-native planted forests, agricultural lands, grasslands, and native forests was sequentially simulated. The hydrological model was initially calibrated for stream flow at the exit of the watershed (R2 = 0.66, P < 0.001). The calibrated model was used for simulating the hydrological responses under two flood events: a peak flood (August 24, 2002), and a moderate flood of (October 14, 2002). The simulated land use scenarios were sequentially and independently input to the model. Results showed that for both flood-events the hydrological response to non-native forests or agricultural lands was very similar to the baseline hydrograph corresponding to 2002 land use. The hydrological response to predominant grasslands and native forests was significantly different. For the August 24 flood, the model calculates that extreme minimum flows and low base flows take place when native forest is the predominant land use. When the watershed is covered by grasslands the most balanced hydrograph (moderate droughts and moderate peak flows) is produced. For the moderate flood event of October 14, the highest values of peak flows and base flows were produced when the watershed is predominantly covered by grasslands. However, the hydrograph for grasslands still showed the most balanced response (moderate droughts and moderate peak flows). For native forests, the response hydrograph shows low base flows but also low peak flows. Therefore, combining grasslands with native forests may provide a balanced hydrograph not only with respect to magnitudes of flow (high base flows and moderated peak flows) but also on their timing. Both conditions are favorable for flood control and drought mitigation.