A new methodology based on the computational intelligence (CI) system is proposed and tested for
modeling the classic 1D dam-break flow problem. The reason to seek for a new solution lies in the
shortcomings of the existing analytical and numerical models. This includes the difficulty of using the
exact solutions and the unwanted fluctuations, which arise in the numerical results. In this research,
the application of the radial-basis-function (RBF) and multi-layer-perceptron (MLP) systems is detailed
for the solution of twenty-nine dam-break scenarios. The models are developed using seven variables,
i.e. the length of the channel, the depths of the up-and downstream sections, time, and distance as the
inputs. Moreover, the depths and velocities of each computational node in the flow domain are considered
as the model outputs. The models are validated against the analytical, and Lax-Wendroff and
MacCormack FDM schemes. The findings indicate that the employed CI models are able to replicate
the overall shape of the shock- and rarefaction-waves. Furthermore, the MLP system outperforms RBF
and the tested numerical schemes. A new monolithic equation is proposed based on the best fitting
model, which can be used as an efficient alternative to the existing piecewise analytic equations.