This study is about fine tuning the targeting capacity of peptide-decorated nanoparticles to discriminate between cells that express different integrin make-ups. Using microfluidic-assisted nanoprecipitation, we have prepared poly(lactic acid-co-glycolic acid) (PLGA) nanoparticles with a PEGylated surface decorated with two different arginine-glycine-aspartic acid (RGD) peptides: one is cyclic (RGDFC) and has specific affinity towards αvβ3 integrin heterodimers; the other is linear (RGDSP) and is reported to bind equally αvβ3 and α5β1. We have then evaluated the nanoparticle internalization in two cell lines with a markedly different integrin fingerprint: ovarian carcinoma A2780 (almost no αvβ3, moderate in α5β1) and glioma U87MG (very high in αvβ3, moderate/high in α5β1). As expected, particles with cyclic RGD were heavily internalized by U87MG (proportional to the peptide content and abrogated by anti-αvβ3) but not by A2780 (same as PEGylated particles). The linear peptide, on the other hand, did not differentiate between the cell lines, and the uptake increase vs. control particles was never higher than 50%, indicating a possible low and unselective affinity for various integrins. The strong preference of U87MG for cyclic (vs. linear) peptide-decorated nanoparticles was shown in 2D culture and further demonstrated in spheroids. Our results demonstrate that targeting specific integrin make-ups is possible and may open the way to more precise treatment, but more efforts need to be devoted to a better understanding of the relation between RGD structure and their integrin-binding capacity.