Computational fluid dynamics (CFD) was used for hydrodynamic modeling of flocculation and sedimentation experiments in a novel spiral hydraulic flocculation/sedimentation system. The experiments were conducted by different flow rates and different internal geometry to obtain different flocculation times and velocity gradients. In all experiments, the velocity gradient was gradually decreased with depth showing very smooth tapered flocculation which is preferred in hydraulic flocculation. The results of velocity gradient that obtained from the CFD simulation were used to investigate the influence of velocity gradient on turbidity and NOM removal in terms of specific ultraviolet absorbance (SUVA) and dissolved organic carbon (DOC). The experimental results revealed high influence of velocity gradient on reduction of NOM and turbidity. Fuzzy inference system (FIS) and response surface methodology (RSM) were used for modeling the influence of initial SUVA, DOC, and turbidity with velocity gradient and flocculation time on the treatment efficiency. Both methods were suitable for describing the treatment process, however, fuzzy inference model validated the experimental results with higher correlation. (C) 2016 Elsevier B.V. All rights reserved.