The purpose of this paper is to studying nonlinear k-ε turbulence models and its advantages in internal combustion engines, since the standard k-ε model is incapable of representing the anisotropy of turbulence intensities and fails to express the Reynolds stresses adequately in rotating flows. Therefore, this model is not only incapable of expressing the anisotropy of turbulence in an engine cylinder, but also is unable to provide good performance when computing the swirling and tumbling flows is important in engine cylinders. Thus, in this paper, the results of nonlinear k-ε model are compared with those of the linear one. Results of diesel engine simulation with linear and nonlinear k-ε models in comparison show that turbulence intensity in the nonlinear model simulation is higher than that of the linear model; also, nonlinear k-ε models predict the second peak value because of the bowl shape in expansion stroke for turbulence intensity. Gas injection results show that nonlinear turbulence models predict spray penetration accurately because of correctly turbulence intensities predicting. Also, the results demonstrate that, for high pressure gas injection, turbulence intensity is high and predicted accurately using nonlinear models. Then, its spray penetration length is predicted accurately in comparison to experimental data’s. Although CPU time spending in the nonlinear model is more than that of the linear one, the non-linear stress model is found to increase computation time by 19%.