To assess flooding risk, hydraulic models are used but they contain many uncertainties related to the lack of knowledge of input parameters. To quantify the uncertainties and evaluate the most influential parameters of the model, Uncertainty Quantification (UQ) and Global Sensitivity Analysis (GSA) can be used. In order to implement these methods, model inputs must be independent, which is not often the case. This research aims to propose a methodology to deal with dépendent inputs in UQ and GSA for hydraulic models and to reduce the computational times. To do so, a 2D hydraulic model of the Loire River built with TELEMAC-2D was used. The study methodology is carried out in the following steps: the uncertain model inputs (hydraulic and breach parameters) and the outputs of interest (the water level at given points) are set. The inputs margins and dependencies are defined by a statistical analysis using a real dataset of the Loire River. The dependency structure between inputs is represented by copulas. Since UQ and GSA require many simulations, kriging metamodels are used to increase the number of experiments in a short time period. Finally, UQ and GSA are carried out by considering the inputs dépendent or not. The outputs distribution slightly differs if the inputs are considered independent or not. The number of influencing parameters increases when inputs are considered dépendent. Some parameters, usually considered as not influencing, may actually be significantly impacting on the outputs. These results suggest that dependencies should not be overlooked in 2D hydraulic models.