Purpose Although bone marrow stimulation (BMS) as a treatment for osteochondral lesions of the talus (OCLT) shows high rates of sport resumption at short-term follow-up, it is unclear whether the sports activity is still possible at longer follow-up. The purpose of this study was, therefore, to evaluate sports activity after arthroscopic BMS at long-term follow-up. Methods Sixty patients included in a previously published randomized-controlled trial were analyzed in the present study. All patients had undergone arthroscopic debridement and BMS for OCLT. Return to sports, level, and type were assessed in the first year post-operative and at final follow-up. Secondary outcome measures were assessed by standardized questionnaires with use of numeric rating scales for pain and satisfaction and the Foot and Ankle Outcome Score (FAOS). Results The mean follow-up was 6.4 years (SD ± 1.1 years). The mean level of activity measured with the AAS was 6.2 pre-injury and 3.4 post-injury. It increased to 5.2 at 1 year after surgery and was 5.8 at final follow-up. At final follow-up, 54 patients (90%) participated in 16 different sports. Thirty-three patients (53%) indicated they returned to play sport at their pre-injury level. Twenty patients (33%) were not able to obtain their pre-injury level of sport because of ankle problems and eight other patients (13%) because of other reasons. Mean NRS for pain during rest was 2.7 pre-operative, 1.1 at 1 year, and 1.0 at final follow-up. Mean NRS during activity changed from 7.9 to 3.7 to 4.4, respectively. The FAOS scores improved at 1 year follow-up, but all subscores significantly decreased at final follow-up. Conclusion At long-term follow-up (mean 6.4 years) after BMS for OCLT, 90% of patients still participate in sports activities, of whom 53% at pre-injury level. The AAS of the patients participating in sports remains similar pre-injury and post-operatively at final follow-up. A decrease over time in clinical outcomes was, however, seen when the follow-up scores at 1 year post-operatively were compared with the final follow-up. Level of evidence Level II.