Abstract: The sintering of agglomerates under high temperature determines their surface area and internal inhomogeneous structure.These available models for numerical simulation of sintering phenomenon generally assume that agglomerate consists of equally-sized primary particles（PP）,which conflicts with experimental measurements.In this paper,it was considered primary particles within agglomerates satisfy polydisperse distribution.The population balance equations for the mass and number of primary particles were constructed to obtain the evolution of size distribution of primary particles within agglomerates.The polydisperse PP model was then used to simulate the sintering process that was used to tailor TiO2 nanoparticles from the gas phase.It is found that the agreement between experimental measurements and numerical results is well in general,and the polydisperse PP model has advantage in numerical precision with respect to the monodisperse PP model（that considers monodisperse PP within agglomerates）.The polydisperse PP model for agglomerate sintering presented in this paper is capable of applying to multiscale modeling,population balance-Monte Carlo simulation,and Eulerian-Lagrangian model for multi-phase flows.