The failure of reinforced concrete structures in recent earthquakes caused concern about the performance of beam column joints. Confinement of joint is one of the ways to improve the performance of beam column joints during earthquakes. This paper describes an experimental study of exterior beam-column joints with two non-conventional reinforcement arrangements. One exterior beam-column joint of a six story building in seismic zone III of India was designed for earthquake loading. The transverse reinforcement of the joint assemblages were detailed as per IS 13920:1993 and IS 456:2000 respectively. The proposed nonconventional reinforcement was provided in the form of diagonal reinforcement on the faces of the joint, as a replacement of stirrups in the joint region for joints detailed as per IS 13920 and as additional reinforcement for joints detailed as per IS 456. These newly proposed detailing have the basic advantage of reducing the reinforcement congestion at the joint region. In order to study and compare the performance of joint with different detailing, four types of one-third scale specimens were cast (two numbers in each type). The main objective of the present study is to investigate the effectiveness of the proposed reinforcement detailing. All the specimens were tested under reverse cyclic loading, with appropriate axial load. From the test results, it was found that the beam-column joint having confining reinforcement as per IS: 456 with nonconventional detailing performed well. Test results indicate that the non-conventionally detailed specimens, Type 2 and Type 4 have an improvement in average ductility of 16% and 119% than their conventionally detailed counter parts (Type1 and Type 3). Further, the joint shear capacity of the Type 2 and Type 4 specimens are improved by 8.4% and 15.6% than the corresponding specimens of Type 1 and Type 3 respectively. The present study proposes a closed form expression to compute the yield and ultimate load of the system. This is accomplished using the theory of statics and the failure pattern observed during testing. Good correlation is found between the theoretical and experimental results.