In many applications, amorphous particles bond together through a phenomenon known as sintering to minimize their surface energy. Water is a plasticizer for many food and pharmaceutical powders and the strong reduction in viscosity induced by moisture absorption can accelerate strongly particle sintering. Numerical simulations of particle sintering usually neglect the coupling with moisture transfer, considering a uniform viscosity throughout the particle. In this study, a novel approach based on solving Navier-Stokes equation using an Arbitrary Lagrangian–Eulerian (ALE) approach is proposed to model the dynamics of particle sintering coupled with moisture transfer. Maltodextrin DE21 is considered as an industrially relevant example of amorphous particles. Due to moisture uptake, strong gradients of viscosity can exist in the particles undergoing sintering. FEM simulations consider accurately the forces acting on the contact area between the particles, leading to slower dynamics than commonly used approximate analytical models. This study highlights that FEM simulations considering a homogeneous moisture and viscosity within the particles are in many cases sufficiently accurate and identifies the limits of validity of this assumption. In the conditions considered in this study, the intraparticle gradients were found to condition significantly the sintering dynamics only when particle diameter is above 1.5mm. The particle size affects strongly both the dynamics of sintering and of moisture transfer. Moreover, higher external relative humidity leads to a lower viscosity and a faster sintering kinetics. The initial water content was found to have a lower impact in the conditions studied. This coupled simulation approach can be used to identify conditions reducing the risk of caking during the storage of amorphous powders or to master sintering during powder structuration processes. Furthermore this study helps identifying when simpler simulation approaches considering homogeneous particles can be safely used and shows the limitations of simplified analytical models.