Recurrent patellar dislocations were treated via knee arthroscopy combined with C-arm fluoroscopy, and reconstruction of the medial patellofemoral ligaments. Between October 2013 and March 2017, 52 cases of recurrent patellar dislocation [27 males and 25 females; age, 16-47 years (mean, 21.90 years)] were treated. Arthroscopic exploration was performed and patellofemoral joint cartilage injuries were repaired. It was subsequently determined whether it was necessary to release the lateral patellofemoral support belt. Pre-operative measurements were used to decide whether tibial tubercle osteotomy was required. Medial patellofemoral ligaments were reconstructed using autologous semitendinosus tendons. Smith and Nephew model 3.5 line anchors were used to double-anchor the medial patellofemoral margin. On the femoral side, the medial patellofemoral ligament was fixed using 7-cm, absorbable, interfacial compression screws. All cases were followed for 1-40 months (average, 21 months). The Q angle, tibial tuberosity trochlear groove distance, Insall-Salvati index, patellofemoral angle, lateral patellofemoral angle and lateral shift were evaluated on X-Ray images using the picture archiving and communication system. Subjective International Knee Documentation Committee (IKDC) knee joint functional scores and Lysholm scores were recorded. Post-operative fear was absent, and no patellar re-dislocation or re-fracture was noted during follow-up. At the end of follow-up, the patellofemoral angle (0.22±4.23°), lateral patellofemoral angle (3.44±1.30°), and lateral shift (0.36+0.14°) differed significantly from the pre-operative values (all, P<0.05). Furthermore, IKDC and Lysholm scores (87.84+3.74 and 87.48+3.35, respectively) differed significantly from the pre-operative values (both, P<0.05). These findings suggest that, in the short term, recurrent patellar dislocation treatment via knee arthroscopy combined with C-arm fluoroscopy and reconstruction of the medial patellofemoral ligament was effective.