In this work, the effects of nanoparticle size, particle volume fraction and pH on the viscosity of silicon dioxide nanocolloidal dispersions are investigated. Both size and pH are found to significantly affect nanocolloid viscosity. Two models are used to study the effect of aggregate structure on the viscosity of the nanocolloidal dispersion. The fractal concept is introduced to describe the irregular and dynamic aggregate structure. The structure of aggregates, which is considered to play an important role in viscosity, is affected by both intermolecular and electrostatic forces. The particle interaction is primarily affected by particle distance and becomes stronger with decreasing particle size and increasing volume fraction. The aggregate structure is also affected by the pH of the solution. Studying the relationship between pH and zeta-potential shows that with the neutralization of charges on the particle surface and decreasing electrical repulsion force, the particle interaction becomes dominated by attractive forces and the aggregates form a more compact structure. In recent years, nanocolloidal dispersions, which are heterogeneous mixtures consisting of very small particles with sizes typically in the order of 1-1000 nm, have attracted much attention for applications related to cooling [1], nanolubricants [2], drug delivery and diagnosis [3]. In the application of nanocolloidal dispersions, the viscosity, which is related to the required pumping power and diffusion properties, plays an important role in delivery systems. The rheological behavior of nanoparticle-fluid mixtures have been investigated experimentally [4][5][6][7]. Newtonian behavior has been observed for very dilute suspensions [4,7] and a considerable amount of effort has been devoted to determining the relationship between the viscosity and volume fraction of suspensions. Yu et al.[8] observed shear-thinning behavior when the volume concentration was higher than 0.03. An interesting phenomenon is that most of the viscosities are under-predicted by the traditional Einstein viscosity model [9,10]. This discrepancy could be caused by aggregation of the nanoparticles and an underestimated effective volume fraction of clusters [10,11]. Rubio-Hernandez et al.[6] also noted that the intrinsic viscosity will be affected by the shape of aggregates, which could be changed by the pH and shear rate of a colloidal system. The formation of aggregates can greatly change the transfer of heat and stress in a system because of their porous structure and large effective volume fraction. Particularly for nanocolloidal dispersions, the distance between aggregates can be in the order of several nanometers. The aggregation of nanoparticles can simply be characterized as ellipsoidal with a certain effective size and axial ratio. The size and shape of aggregates can be determined from dynamic scattering light measurements and intrinsic viscosity data, respectively [6]. Aggregation is a dynamic process and the structure changes continuously because of Brownian m...