The response of unreinforced beam-column joints represents a key issue in seismic vulnerability and fragility assessment of existing Reinforced Concrete (RC) buildings. A great attention has been focused on the investigation of the experimental response of unreinforced joints and their possible strengthening strategies.In this study, an experimental investigation of the response of exterior RC beam-column joints strengthened with prestressed steel strips is presented. Four full-scale specimens were tested, two non-strengthened and two strengthened. The non-strengthened specimens were built with (specimen S) or without (NS) stirrups in the joint panel and were tested for reference and comparison purposes, in order to reproduce a seismic code-compliant and a non-conforming RC joint, respectively. The strengthened specimens had two different layouts characterized by a higher (CAM1) or lower (CAM2) complexity of installation and level of building disruption.The response of specimen NS was controlled by joint panel failure following beam yielding, whereas the specimen S showed a ductile response, controlled by beam flexural failure. The response of specimens CAM1 and CAM2 demonstrated the effectiveness of the adopted technology. For CAM1, a ductile response was observed, similar to specimen S. For CAM2, a joint panel failure following beam yielding was observed, as for specimen NS, but with a significantly higher global ductility.The presence of prestressed steel strips led to a delay in the onset of diagonal cracking, with a slight increase in strength at this condition, i.e. by 8%, but a significantly higher deformation capacity, with the drift increasing, on average, by 122%. The strengthening led to a relatively slight increase in the resistance of the subassemblage, which was very similar in specimens S, CAM1 and CAM2, and, on average, 11% higher than in specimen NS. Instead, a much more significant difference was observed in terms of deformation capacity, with the drift at conventional ultimate condition in specimens CAM1 and CAM2 77% and 61% higher than specimen NS, respectively.The experimental results are reported, and the effectiveness of the tested strengthening solutions is discussed, in terms of global response of the subassemblage, observed damage evolution, energy dissipation capacity, and local response, such as joint shear distortion and measured strain of strengthening steel strips.