The two highly homologous non-visual arrestins, beta-arrestin 1 and 2, are ubiquitously expressed in the central nervous system, yet knowledge of their disparate roles is limited. While beta-arrestin 2 (βarr2) has been implicated in several aspects of reward-related learning and behavior, very little is known about the behavioral function of beta-arrestin 1 (βarr1). Using mice lacking βarr1, we focused on the role of this scaffolding and signal transduction protein in reward-motivated behaviors and in striatal glutamatergic function. We found that βarr1 KO mice were both slower in acquiring cocaine self-administration and in extinguishing this behavior. They also showed deficits in learning tasks supported by a natural food reward, suggesting a general alteration in reward processing. We then examined glutamatergic synaptic strength in WT and KO medium spiny neurons (MSNs) of the Nucleus Accumbens (NAc) shell in naïve animals, and from those that underwent cocaine self-administration. An increase in the AMPA/NMDA (A/N) ratio and a relative lack of GluN2B-enriched NMDARs was found in naïve KO vs WT MSNs. Applying Lim Domain Kinase (LIMK1), the kinase that phosphorylates and inactivates cofilin, to these cells, showed that both βarr1 and LIMK regulate the A/N ratio and GluN2B-NMDARs. Cocaine self-administration increased the A/N ratio and GluN2B-NMDARs in WT MSNs and, although the A/N ratio also increased in KO MSNs, this was accompanied by fewer GluN2B-NMDARs and an appearance of calcium-permeable AMPARs. Finally, to examine the consequences of reduced basal GluN2B-NMDARs in reward-processing seen in KO mice, we chronically infused ifenprodil, a GluN2B antagonist, into the NAc shell of WT mice. This intervention substantially reduced food-motivated behavior. Together these findings identify a previously unknown role of βarr1 in regulating specific reward-motivated behaviors and glutamatergic function.