The paper discusses the state-of-the-art of turbulence modeling and presents possible ways for improvement in the future turbulence models. Based on a set of turbulence closure postulations, a variation of second-order turbulence models, such as the Reynolds stress model (RSM), the algebraic stress model (k-ε-A), and the eddy viscosity model (k-ε-E), are obtained. Examples of prediction made are free shear flows, cavity flows, and flows past an off-set channel. Although a complete turbulence model does not exist at the present time, some prediction capability has been achieved by the secondorder turbulence model. The incompleteness of turbulence modeling may be attributed to the inadequacy of isotropic dissipation and single turbulent scale postulations. Use of multiple turbulence scale concepts, including use of fractal dimension of turbulent eddies may improve turbulence prediction.