Abstract: Secondary flow has a substantial influence on water flow, mass transportation and riverbed evolution. However, conventional 2-D depth-averaged models usually ignore the influence of secondary flow. Furthermore, the suitability of existing secondary flow correction models is debatable. To select a secondary flow correction method that can be applied to complex consecutive bends, two typical linear model calculation schemes, the Delft3D and Lien methods, are evaluated. The influence of secondary flow is considered in the 2-D depth-averaged models proposed in this paper by calculating the dispersion stresses added to the 2-D shallow water momentum equations in orthogonal curvilinear coordinates. Four unique consecutive curved flume experiments were chosen to evaluate the simulation effects. The evaluations compared the simulated water levels and velocity distributions for a conventional 2-D model and two kinds of secondary flow correction methods. Analysis of the simulation results demonstrates that the performance of the Delft3D secondary flow correction method is the better of the two methods tested. The method has higher degrees of freedom and is suitable for different complexity levels of consecutive bend simulations. While Lien's secondary flow correction method is not suitable for sharp consecutive bend flow simulations, it can be used effectively to simulate weak or moderate bends, however, care should be taken for consecutive bends of moderate curvature. Therefore, comparison of the two methods shows that the Delft3D modeling method is the better choice for simulating flow in consecutive bends.