This retrospective study is aimed to evaluate the long-term outcome of arthrocentesis for the treatment of temporomandibular joint (TMJ) internal derangements, and to document the impact of patient, time and symptom-related factors on the outcome of the procedure. Thirty-four patients (48 joints) who underwent TMJ arthrocentesis were included in this study. The scores for preoperative maximal mouth opening, and VAS scores for pain and dysfunction were compared with the follow-up scores obtained by the questionnaire and clinical examination. Several factors that may affect the long-term outcome of arthrocentesis are further evaluated. Statistical evaluation of the baseline and follow-up data was made by linear regression analysis and paired t-test. The mean follow-up period was 22 months. There was a significant (P<0.001) increase in the maximal mouth opening (MMO) postoperatively that held during the longer term follow-up period. The pain and dysfunction levels at the follow-up were significantly (P<0.001) lower than the preoperative values. Twenty-six per cent of patients were pain free and also showed a total relief in dysfunction at the follow-up. Although both preoperative and follow-up pain scores were higher in patients with bruxism, there was not a significant difference in the outcome when compared with non-bruxers. However, there was a greater reduction in dysfunction with improvement in MMO in non-bruxers. The duration of symptoms before arthrocentesis has not been found to affect the outcome. Also, there were no significant differences between the results of follow-up when comparing the shorter follow-up time results (<20 months) and longer term results. Arthrocentesis for the treatment of TMJ internal derangements offers favourable long-term stable results with regard to increasing maximal mouth opening, and reducing pain and dysfunction.