Nanofluid can enhance heat transfer due to the suspending nanoparticles. The mechanism of heat transportation by nanoparticles remains unclear so far. Aggregation of nanoparticles, one of the important mechanisms to elevate the thermal conductivity of nanofluid, was proved by not a few researches. However, the aggregation morphology of nanoparticles evaluated by fractal dimension will greatly influence the thermal conductivity of nanofluid. In this paper, equilibrium molecular dynamics simulations were carried out to calculate the thermal conductivity via Green-Kubo formula. In contemporary, fractal dimensions of the aggregations with various morphologies were obtained by Schmidt-Ott equation. Comparisons of the fractal dimension and thermal conductivity of the nanofluid with same volume fraction show us that, lower fractal dimension can deduce greater thermal conductivity. In addition, the difference of 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. This is helpful for us to understand the influence of aggregation morphology of nanoparticles on the thermal conductivity of nanofluid. (C) 2018 Elsevier Ltd. All rights reserved.