Floods are a common natural disaster that causes huge economic, social and human losses. Of several methods of flood routing, the dynamic wave model is one of the best approaches for predicting the characteristics of flooding during its propagation in natural rivers because all the terms of the moment equation are considered in the model. However, no significant research has been done on how the sensitivity of the model affects the accuracy of the hydro-graph downstream. In this study, a comprehensive analysis of the input parameters of the wave dynamic model was carried out using field applications in natural rivers and routing experiments in artificial channels using the multi-parametric sensitivity analysis (GMPSA). The results indicate that the effects of errors of the input parameters on the output results are more significant in special situations, such as lower values of the Manning coefficient of roughness and / or a more pronounced slope of the bed on the characteristics of a hydro-graph of design, greater values of the asymmetry factor and / or time to peak in the channel characteristics, higher values of the Manning roughness coefficient and / or bed slope in the space passage, and lower values of the Manning roughness coefficient and / or a steeper slope of the bed in the time step and weighting factor.