The CFD method, as the initial analysis has more benefits regarding experiments, including saving time and costs. Variable of flow parameters and geometry can be developed easily to get the desired results. However, research is needed to improve the accuracy of the results and the efficiency of the calculation process. The study of complex turbulent flow modeling becomes very important. The k-ϵ model, and renormalization group (RNG) k-ϵ model are widely used in research, to produce the appropriate models and develop the value of constants. This turbulent flow modeling research was conducted to improve the accuracy of the results and the efficiency of the calculation process in turbulent flow of crossflow turbine. Research is done by comparing the simulation results of k-ϵ model and RNG k-ϵ model. The comparison of the k-ϵ model and RNG k-ϵ model results shows different results for predicting the average pressure and velocity distribution in turbulent flow of crossflow turbine. Likewise for turbulent parameters. It can be concluded for complex fluid flow recommending RNG k-ϵ model.