Abstarct: Heat removing and temperature control are critical in various industries such as electronics, aerospace, military with high production of heat flux. In the recent decade, efforts have been concentrated on utilizing nanofluids as coolant fluids. Using the flow of these fluids could improve the efficiency of micro heat sinks. In the present research, the numerical and experimental studies of convection heat transfer of nanofluid flows through a microchannel have been performed. Therefore, a numerical code has been developed by solving two dimensional elliptical governing equations using finite volume method. The numerical code has been used to investigate the brinkman number, substrate thickness and material. Results of the numerical calculation show good agreement with the experimental and analytical analysis in the literature. The result shows that an increase in wall thickness results in thermal variation in the solid- fluid interface. It arises from the so-called axial conduction through the microchannel substrate. Wall thickness in numerical calculations cause decrease in convection heat transfer coefficient. increasing in the volume fraction of nanoparticles enhances the convection heat transfer coefficient. Therefore, the axial variation of heat flux would be more uniform at the solid-fluid interface as a result of decreases in axial conduction. Utilizing high conductivity substrate decreases the maximum temperature and thermal resistance of substrate as a result of axial conduction. On the other hand, an integrated experimental set up has been assembled to investigate the hydrodynamics and thermal parameters of nanofluid flow containing single and multi wall carbon nanotubes. The experimental set up consists of various sections in order to drive the nanofluid through 16 parallel microchannels fabricated on the copper substrate.The Nanofluids has been fabricated with dispersing functionalized carbon nanotubes in research institute of petroleum industry. Effective viscosity and thermal conductivity of carbon structured nanofluid have been modeled at the certain range of temperature and weight fraction. Experimental data shows that a flow viscosity arisen from surface roughness plus effective viscosity of nanfluid relate an inverse relation between non-dimension length and poissle number at a fully developed region. However, Surface roughnesses have no effect on the poissle number at low Reynolds number. Therefore, flows through rough microchannel correspond to the flows through smooth one. An increase in Grashof number causes an enhanced buoyancy forces and therefore intensify local fluctuations near the rough surface. In general, although inclusion of nanoparticles increases convection heat transfer coefficient at low Reynolds number, weight fraction and particle type have no significant effect on the heat transfer coeeficient. On the other hand, for higher Reynolds number ( ),weight fraction and particle type affect the heat transfer coefficient of nanofluid flow.