No comparative evaluations of anti-dyskinetic drugs to mitigate levodopa-induced dyskinesia (LID) and OFF time have been performed; head-to-head anti-dyskinetic trials are unlikely to be funded. We sought to compare the efficacy, tolerability, and safety of 12 anti-dyskinetic drugs for peak-dose dyskinesia and OFF time through a Bayesian network meta-analysis.A network meta-analysis (NMA) was conducted for randomized controlled trials (RCTs) searched through the databases of MEDLINE, Embase, Cochrane Controlled Register of Controlled Trials from incipiency to January 2025. The main endpoints were the Unified Dyskinesia Rating Scale (UDysRS), modified Abnormal Involuntary Movement Scale, or the Unified Parkinson's Disease Rating Scale IV score for LID and the number of hours of OFF time in a standardized OFF-ON diary for OFF time reduction.Twenty-one RCTs were included in the analysis. The surface under the cumulative ranking curve (SUCRA) indicated that amantadine extended-release (ER) was associated with the largest improvement in the UDysRS (0.934), followed by amantadine immediate-release (IR) (0.680). For OFF time reduction, amantadine IR (0.902) and amantadine ER (0.833) were the two best treatments compared with foliglurax (0.526), mavoglurant (0.464) and sarizotan (0.012). Fipamezole ranked first in overall withdrawals (0.812). Amantadine ER ranked first in withdrawals due to adverse events (AEs) and the incidences of AEs (≥ 1 AEs) (0.900 and 0.887).Amantadine ER improved dyskinesia significantly versus amantadine IR. Both formulations of amantadine were superior as OFF time-reducing agents compared to other anti-dyskinetic drugs, whereas amantadine IR non-significantly reduced OFF time versus amantadine ER.© 2025 The Author(s). European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology.