OBJECTIVE To determine the complication rate for ankle arthroscopy. METHODS A review of a consecutive series of patients undergoing ankle arthroscopy in our hospital between 1987 and 2006 was undertaken. Anterior ankle arthroscopy was performed by means of a 2-portal dorsiflexion method with intermittent soft tissue distraction. Posterior ankle arthroscopy was performed by means of a two-portal hindfoot approach. Complications were registered in a prospective national registration system. Apart from this complication registry, patient records, outpatient charts and operative reports were reviewed. Patients with a complication were asked to visit our hospital for clinical examination and assessment of permanent damage and persisting complaints. RESULTS An overall complication rate of 3.5% in 1,305 procedures was found. Neurological complications (1.9%) were related to portal placement. Age was a significant risk factor for the occurrence of complications. Most complications were transient and resolved within 6 months. Complications did not lead to functional limitations. Residual complaints did not influence daily activities. CONCLUSIONS Our complication rate is less than half of what has been reported in literature (3.5 vs 10.3%). The use of the dorsiflexion method for anterior ankle arthroscopy can prevent a significant number of complications. Posterior ankle arthroscopy by means of a two-portal hindfoot approach is a safe procedure with a complication rate that compares favourably to that of anterior ankle arthroscopy. METHODS Retrospective prognostic study, Level II.