Objective Treatments for steroid-induced osteonecrosis of the knee remains challenging, and there has not been sufficient evidence to support joint preservation surgery. This study evaluated long-term outcomes of osteochondral autologous transplantation (OAT) for steroid-induced osteonecrosis of the knee. Design This retrospective case series included patients who underwent OAT for steroid-induced osteonecrosis of the knee from 1998 to 2008. The survivorship and need for secondary surgery were evaluated, and the clinical outcome was evaluated with the International Knee Documentation Committee (IKDC) subjective score. Preoperative and final Kellgren-Lawrence (KL) grade of the femorotibial and patellofemoral joints were individually evaluated. Results Fourteen knees of 10 patients whose mean age was 32.5 (95%CI 26.4-38.6) years were included and followed for 14.0 (12.4-15.7) years. The mean lesion size of 6.9 (5.3-8.5) cm2 was repaired using 4 median (minimum 2, maximum 5) osteochondral plugs. No revision surgeries were performed for transplanted osteochondral plugs. The IKDC subjective score improved from 32.9 (24.5-41.3) to 74.2 (61.9-88.5) ( P < 0.001). Knee flexion was improved at the final follow-up, and Seiza sitting was finally possible in 9 knees in 7 patients. Although the osteoarthritic change did not progress in femorotibial joint, patellofemoral joint showed early osteoarthritic changes at the final follow-up (mean KL grade: 0.8 [0.5-1.1]). Conclusions Prosthetic joint replacement was successfully avoided for at least the first decade by OAT in young patients with steroid-induced osteonecrosis of the knee. The progression of KL grade of the patellofemoral joint is of concern.