The research presented in this Thesis aims at defining the strengths and weaknesses of an Improved 1D/1D model when compared with a more accurate 1D/2D model. Although both coupled-models (sewer/surface) solve the St.\ Venant equations in both layers, the latter uses a higher approximation (2D two-dimensional) on the surface layer. Consequently, the 1D/1D model is computationally more efficient when compared to the 1D/2D model, however there is some compromise with the overall accuracy. The hypothesis is that "The inundation extent of urban flooding can be reproduced by 1D/1D models in good agreement with the 1D/2D models if the results are kept within certain limits of resolution and under certain conditions". The Thesis starts by investigating ways of improving an existing 1D/1D model to rival the more accurate 1D/2D model. Parts of the 1D/1D model code are changed and new algorithms and routines implemented. An innovative GIS tool translates the 1D output-results into 2D flood-inundation-maps enabling a thorough comparison between the two models. The methodology assures the set-up of two equivalent models, which includes a novel algorithm for calibrating the 1D/1D model vs.\ the 1D/2D model results. Developments are tested in two distinctly different case studies of areas prone to flooding. The conclusion is that the 1D/1D model is able to simulate flooding in good agreement with the 1D/2D model; however, it is found that features such as topography, density of the urbanised areas and rainfall distribution may affect the agreement between both models. The work presented herein is a step forward in understanding the modelling capabilities of the analysed coupled-models, and to some extent may be extrapolated to other models. Research is growing in urban flooding and this work may well prove to be a strong foundation basis for future research.