PurposeThis study aims to research the prediction performance of the bifurcation approach with different base models in different kinds of turbulent flows with rotation and curvature.Design/methodology/approachThe k−ω and Shear-Stress Transport (SST) k−ω models are modified by using the complete eddy viscosity coefficient expression, and the latter is modified by using two sets of model coefficients. The two bifurcation models were tested in three cases: rotating channel flow with system rotation, Taylor–Couette flow with wall rotation and curvature effect and swirling flow through an abrupt axisymmetric expansion with inlet swirling flow.FindingsIn these flows, the bifurcation approach can significantly improve the prediction performance of the base model in the fluctuation velocity. The deviation of the BSkO model is slightly superior to the BkO model by about 2% in the Taylor–Couette flow. The prediction effect of the root-mean-square (RMS) velocity of the BSkO model increases by about 4–5% as the number of grids increases about 2.37 times, and the best is the Large Eddy Simulation (LES) grid used. Finally, compared with the SST k−ω model, the average iteration time of the SST with curvature correction (SST-CC), bifurcation k−ω (BkO) and bifurcation SST k−ω (BSkO) models increased by 27.7%, 86.9% and 62.3%, respectively.Originality/valueThis study is helpful to understand further the application of the bifurcation method in the turbulence model.