The estimation of flow in river basins is frequently required in hydrological practice and is of great economic significance. This study uses streamflow and catchment characteristic data, including land use information, from the Sperchios river basin (Central Greece) and applies quantile regression modelling. From 100% of the observations, 20% were selected randomly and used for independent testing of the quantile regression. The test observations and observations used for the model development were found to have similar characteristics. The quantile regression method generally provides accurate flow estimates; however, these techniques involve a chance of large errors under particular conditions, and provisions should be made accordingly.