A computational model combining the fluid dynamics with the particle kinetics was employed to study TiO2 nanoparticle synthesis in a diffusion flame reactor. A one-step chemical kinetics approach was used to model titanium tetraisopropoxide (TTIP) decomposition that leads to homogeneous nucleation and particle formation. The momentum, heat, and mass transfer, Brownian coagulation and diffusion, surface growth, coalescence and thermophoresis have been taken into account. Based on the particle size distributions, an efficient quadrature method of moments was allowed to approximate the general dynamics equation of particle, and the eddy dissipation concept (EDC) combustion model was used to estimate the flame temperature field. Excellent agreements between the model predictions and experimental data, with respect to the flame temperature distribution and particle kinetics, are reached. By taking the particle size and surface area as independent variables, the full distributions of volume equivalent diameters, evolution of the agglomerate number, the geometric standard deviation based on volume and agglomerate fractal nature, mean primary particle size and the number of primary particles per agglomerate are revealed. The variation of oxygen flow rate has an important influence on the temperature distribution and hence on the particle kinetics accordingly. (c) 2007 Published by Elsevier B.V.