This study investigates the Proper Orthogonal Decomposition (POD) combined with Radial Basis Functions (RBF) and Response Surface Method (RSM) based metamodels to appreciate the influence of the system parameters on elastically-mounted test cylinder (TC) behavior excited by water cross-flow. The TC mass, stiffness and damping coefficients and excitation forces are chosen as the design parameters. System responses at selected training points are obtained by full, two-dimensional model simulations using Computational Fluid Dynamics. Obtained results show that POD-RBF metamodel has good accuracy to predict the system parameters, which is not time-dependent, whereas the RSM metamodel can reconstruct the time-dependent TC movement trajectory with high precision.