Nano-remediation is a promising in situ remediation technology. It consists in injecting reactive nanoparticles (NPs) into the subsurface for the displacement or the degradation of contaminants. However, due to the poor mobility control of the reactive nanoparticle suspension, the application of nano-remediation has some major challenges, such as high mobility of the particles, which may favor override of the contamination, and particle aggregation, which can lead to a limited distance of influence. Previous experimental studies show the potential of combining nano-remediation with foam flooding to overcome these issues. However, in order to design and optimize the process, a model which couples nanoparticle and foam transport is necessary. In this paper, a mechanistic model to describe the transport of NPs with and by a foam is presented. The model considers the delivery of nanoscale zero-valent iron (nZVI) and accounts for the processes of aggregation, attachment/detachment, and generation/destruction. Simulations show that when NPs are dispersed in the liquid phase, even in the presence of a foam, they may travel much slower than the NPs carried by the foam bubbles. This is because the nanoparticles in suspension are affected by the attachment onto the rock walls and straining at the pore throats. When the nanoparticle surface is, instead, modified in order to favor their adsorption onto the gas bubbles, NPs are carried by the foam without retardation, except for the small fraction suspended in the liquid phase. Moreover, very stable high-quality foam ( i.e., 80-90 vol% of gas, can be attained using properly surface-modified nZVI (i.e., a nanoparticle-stabilized foam), allowing a significant reduction of water for the operation, while increasing the efficiency of nZVI delivery, even in a low-permeability medium within the shallow subsurface.