In recent years, urban inundation modelling has been stimulated by technological advancements and an increase in data availability and quality. Diminution of modelling restrictions allowed for detailed methodologies, yet escalated the number of required parameters, leading to higher complexity and predictions’ uncertainty issues. In the present study, two state-of-the-art models with distinct numerical discretisation strategies, schematisation and rainfall runoff modules, namely MIKE FLOOD and D-Flow FM are evaluated in urban fluvial and pluvial inundation. The case study of Colombo Metropolitan region (Sri Lanka) is used. Dominant factors interpreting models’ differentiations are identified and further analysed. Results suggest that D-Flow FM concept is more realistic and sensitive to overland surface information but require higher detail and computational times compared to MIKE FLOOD methodology. The outcome of this research is anticipated to aid modellers in choosing the most appropriate model principle based on the data availability, desired accuracy and computational time.