Closure problem remains one of the challenging issues in the parameterization of turbulent fluxes due to the restriction of available physical theory and lack of observation in the parameterization of planetary boundary layer (PBL) processes for several decades. A level 2.5 scheme proposed by Mellor and Yamada is a compromise between locally determined first-order scheme and computationally expensive second order scheme. The scheme evaluates the eddy viscosity coefficient based on the predicted turbulent kinetic energy (TKE). Recent study also makes use of predicted eddy dissipation rate (ε) of mixing length equation-ε so called E-ε model. The model has been successfully applied to the mesoscale simulation and atmospheric boundary layer problems. Motivated by this, a prognostic equation for the eddy dissipation rate is added to one level 2.5 scheme (G-S; Gayno et al 1994) used in 1-D MM5 model in this study. Then the modified model is employed to the homogeneous observation cases from the First ISLSCP (International Satellite Land Surface Climatology Project) Field Experiment (FIFE). Preliminary results suggest that both schemes reproduce similar wind structure at the early stage of PBL development. When boundary layer becomes more convectively unstable, the upper level wind with E-ε scheme agrees well with the observation than the original G-S scheme. In addition, E-ε scheme generates weak TKE in the upper boundary layer which is caused by the small net production of TKE and the eddy dissipation. Further work needs to consider more simulated turbulent boundary layer properties and explore the role of eddy dissipation plays in the development of the PBL structure.