Prediction of colloidal nanoparticle aggregation is an important problem which needs to be solved in an accurate and efficient manner. 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 ggregation of colloidal nanoparticles. Because the BD technique is essentially a reduction of the MD method the accuracy requirements for BD simulations have been established. A new method to match aggregation statistics obtained from MD and BD simulations is proposed. In this method the evolution of the second-order density for MD model is derived. The average relative acceleration between nanopartilce pairs is identified as an important link between MD and coarse-grain simulations such as BD.