Natural biological materials usually adopt functional gradient designs within interfacial regions to fulfil unusual mechanically-challenging demands. Manufacturing analogous gradients to alleviate premature failures for synthetic interfaces has remained challenging until recently, where magnetically-actuated gradient nanocomposites (MA-G-NCs) have emerged as a promising processing technique. The essence of this technique lies in controlling the spatial distribution of nanoreinforcements (usually particles) inside a polymer matrix through a magnetophoresis process. Herein, we present a theory-experiment-combined study on the evolution kinetics and equilibrium distribution of the nanoparticles during the magnetophoresis process and consequently to explore the spatial and temporal tunability of the MA-G-NCs. Using a simplified drift-diffusion theory as the guide, we determine two critical processing parameters for the MA-G-NCs: the applied magnetic field and the actuation duration. By systematically varying these two parameters independently, we experimentally demonstrate that the profile of the nanoparticle distribution inside the MA-G-NCs can be finely tuned both spatially and temporally. In order to quantify the volume fraction of the nanoparticles along the cross section of the MA-G-NCs, we propose a mechanics-based method by site-specifically measuring the local elastic modulus and converting back to the volume fractions based on an established modulus-fraction correlation. The nanoparticle concentration profiles obtained thereby are validated by morphological characterizations and also agree well with theoretical predictions based on the drift-diffusion theory. Our combined results indicate that the magnetophoresis-induced evolution of the nanoparticles follows approximately the drift-diffusion transport process and the gradient profile of the MA-G-NCs is highly controllable and programmable. The presented study not only advances the fundamental understanding of the evolution kinetics of the nanoparticles under the effect of magnetophoresis, but also establishes the critical processing-structure-property relationships for the MA-G-NCs that should guide future development of customized interfaces with desired mechanical and physical property gradients.