This study is concerned with the critical evaluation of predicative capability of a k-$varepsilon$ turbulence model using the Renormalization Group(RNG) theory. The present numerical model for solution of the Navier-Stokes System is based on the modified PISO algorithms. Computations have been performed with the RNG-based K-$varepsilon$ model for the two-dimensional flow over a backward-facing step, a confined coaxial jet, and a swirling flow in a swirl combustor. Numerical results are compared with experimental data in terms of mean flow velocities, turbulent kinetic energy, and turbulent stresses. Numerical results clearly indicate that the RNG-based K-$varepsilon$ turbulence model shows a significant improvement over a standard K-$varepsilon$ model in predicting the turbulent flows with flow separation and swirl.