Prediction of colloidal nanoparticle aggregation is an important problem which needs to be solved in an accurate and efficient manner. Thus, accurate modeling of physico-chemical interactions is required. In addition, to have good aggregation statistics, large system sizes should be considered, which can significantly increase computational cost and decrease code efficiency. In ideal case model which is chosen to predict colloidal nanoparticle aggregation should accurately describe physico-chemical interactions of relatively large physical systems, and at the same time, simulate at low computational cost.;In this research, two simulation approaches, molecular dynamics (MD) and Brownian dynamics (BD), are analyzed and compared with a view to accurately predicting aggregation of colloidal nanoparticles. Based on this comparison it was found that MD approach is not feasible to simulate an aggregation of colloidal nanoparticles for systems of physical sizes but can be used to simulate non-physical (model) systems. Thus, such coarse-grain approaches as BD have to be used instead. Because the BD technique is essentially a reduction of the MD method the accuracy requirements for BD simulations have been established.;Analysis of characteristic time scales for the BD approach justifies reduction of position and velocity-Langevin equations to position-Langevin for physical system and is not justified for model system.;A new method to match aggregation statistics obtained from MD and BD simulations is proposed in this work. In this method the evolution of the second-order density for MD model is derived. The average relative acceleration between nanoparticle pairs is identified as an important link between MD and coarse-grain simulations such as BD.;Although BD is a coarse-grain model with fewer degrees of freedom, it gives reasonable predictions of nanoparticles aggregation. Also, because of its high computational efficiency this method can be a useful tool to simulate nanoparticle aggregation in colloidal systems.