The great interest of many researchers has been aroused in recent two decades due to the great heat transfer enhancement of nanofluid as a heat transfer medium. The reason why the nanofluid can enhance heat transfer is that a number of nanoparticles are suspended in the carry fluid. Most of researchers believe that the microconvection induced by Brownian motion of nanoparticle, the nanolayer around the nanoparticle, the aggregation of nanoparticles and near-field radiation are the underlying mechanisms for heat transfer enhancement by nanofluid. However, contradictories and inconsistencies among experimental results, theoretical results and numerical results are existent commonly because the mechanism of heat transportation by nanoparticles remains unclear so far. Quite a few researches have proven that the aggregation of nanoparticles is one of the important mechanisms for elevating the effective thermal conductivity (ETC) of nanofluid. However, the aggregation morphology (AM) of nanoparticles evaluated by fractal dimension (FD) will greatly influence the thermal conductivity of nanofluid. Unfortunately, all of the existing ETC models are based on the effective medium theory under the assumption of " static state” and " homo-dispersion”. In the present work, equilibrium molecular dynamics (EMD) simulations are carried out to calculate the thermal conductivity of Cu-Ar nanofluid via Green-Kubo formula. In existing researches, fractal dimensions of the aggregations with various morphologies are obtained by Schmidt-Ott equation. Comparisons between the ETC and FD of the nanofluid with same volume fraction show that lower FD can possess greater ETC. It is the first time that the quantitative relationship between ETC and FD has been analyzed. In addition, the difference between loose and compact aggregation can be read out of the pair correlation function near nanoparticles. And the solvent atoms in nanolayer are mobilized and dynamically balanced. The results obtained in the present research are conducible to understanding the influence of AM of nanoparticles on the ETC of nanofluid.