One of the most important industrial aims of researchers and engineers is energy efficiency and increasing yield. Due to the unique features, heat pipe has a special position in systems to optimize energy consumption and energy recovery. Heat pipe is an efficient appliance to transfer heat in a limited ambient. One of the limited factors of heat transfer ability in pipes is the properties of working fluids. In order to overcome this limitation, thermo-physical properties of the fluid can be improved. One of the new ideas to increase the thermal conductivity constant of liquid is distribution of nanoparticles in the main fluid. This new generation of conductive fluids with nanoparticles are referred to as nanofluids. The nanoparticles within the fluid change its thermal conductivity, viscosity and density. So some researchers are trying to apply nanofluids to heat pipes for enhancing their heat transfer performance. In this investigation, computational fluid dynamic (CFD) was used to modeling of fluid and two dimensional heat transfer with phase change in a heat pipe thermosypho in which distilled water and nanoparticle were working fluids. Modeling was done by the software “fluent 12” and desired geometry was made by “gambit”. In the first stage, copper oxide nanofluid at a concentration of 1.0 wt % in basic fluid of distilled water was used as the working fluid. Modeling of two phase fluid was done with both the default model of the software and also coding procedure with considering the effects of non-condensation gases. Convection in thermosyphon and condensation and evaporation was simulated by calculations. Temperature distribution and heat transfer coefficients in CFD were compared with the experimental results, high compatibility was achieved. Results show that in the existence of nanofluids, temperature distribution was more uniform. Also it was clear that by using nanofluid water/ copper oxide instead of distilled water in thermosyphon results in increasing heat transfer coefficients. For instance, heat transfer coefficient of nanofluid evaporation in heat load of 80 W was 72% higher than distilled water. The effect of the presence of non-condensation gases was studied by changes in heat transfer coefficient of condensation. These gases reduced the heat transfer coefficient of condensation to numbers lower than experimental values. Also, the effect of the presence of nanofluids on the total thermal resistance was investigated. In equal input powers, the reduction in total thermal resistance of nanofluid was observed in comparison with distilled water. Then in the second stage, effect of kind and concentration of nanofluids was investigated. In order to study the effect of nanofluid, modeling was done in concentrations 0.1 and 0.5 wt %. By increase in nanofluid concentration from 0.1 to 1 percent, increase of 68% in heat transfer coefficients was observed. Finally, the copper oxide nanofluid was substituted with gold nanofluid 1.0 wt % and the new results are compared with copper oxide nanofluid results. These results shows that evaporation heat transfer coefficient of gold nanofluids was higher than copper oxide nanofluids which was more clear in higher input power. Key Words : Thermosyphon, evaporation, condensation, computational fluid dynamic