An ensemble-modeling scheme incorporating coarse-grained simulations with experimental small-angle X-ray scattering (SAXS) data is applied to dengue virus 2 (DENV2) nonstructural protein 5 (NS5). NS5 serves a key role in viral replication through its two domains that are connected by a 10-residue polypeptide segment. A set of representative structures is generated from a simulated structure pool using SAXS data fitting by the non-negativity least squares (NNLS) or standard ensemble optimization method (EOM) based on a genetic algorithm (GA). It is found that a proper low-energy threshold of the structure pool is necessary to produce a conformational ensemble of two representative structures by both NNLS and GA that agrees well with the experimental SAXS profile. The stability of the constructed ensemble is validated also by molecular dynamics simulations with an all-atom force field. The constructed ensemble successfully revealed the domain-domain orientation and domain-contacting interface of DENV2 NS5. Using experimental data fitting and additional investigations with synthesized data, it is found that energy restraint on the conformational pool is necessary to avoid overinterpretation of experimental data by spurious conformational representations.