Drug efficacy generally varies with different durations. There is no systematic review analyzing the effect of selegiline for Parkinson's disease (PD) on different treatment duration. This study aims to analyze how the efficacy and safety of selegiline changes for PD over time.PubMed, the Cochrane Library, Embase, China National Knowledge Infrastructure and Wanfang Database were systematically retrieved for randomized controlled trials (RCTs) and observational studies of selegiline for PD. The search period was from inception to January 18th, 2022. The efficacy outcomes were measured by the mean change from baseline in the total and sub Unified Parkinson's Disease Rating Scale (UPDRS), Hamilton Depression Rating Scale (HAMD) and Webster Rating Scale (WRS) scores. The safety outcomes were measured by the proportion of participants having any adverse events overall and that in different system organ classes.Among the 3,786 studies obtained, 27 RCTs and 11 observational studies met the inclusion criteria. Twenty-three studies reported an outcome which was also reported in at least one other study, and were included in meta-analyses. Compared with placebo, selegiline was found with a stronger reduction of total UPDRS score with increasing treatment duration [mean difference and 95% CIs in 1 month: -3.56 (-6.67, -0.45); 3 months: -3.32 (-3.75, -2.89); 6 months: -7.46 (-12.60, -2.32); 12 months: -5.07 (-6.74, -3.41); 48 months: -8.78 (-13.75, -3.80); 60 months: -11.06 (-16.19, -5.94)]. A similar trend was also found from the point estimates in UPDRS I, II, III, HAMD and WRS score. The results of observational studies on efficacy were not entirely consistent. As for safety, compared with placebo, selegiline had higher risk of incurring any adverse events [rate: 54.7% vs. 62.1%; odd ratio and 95% CIs: 1.58 (1.02, 2.44)], with the excess adverse events mainly manifested as neuropsychiatric disorders [26.7% vs. 31.6%; 1.36 (1.06, 1.75)] and no significant change over time. The statistically difference in overall adverse event between selegiline and active controls was not found.Selegiline was effective in improving total UPDRS score with increasing treatment duration, and had a higher risk of incurring adverse events, especially the adverse events in the neuropsychiatric system.https://www.crd.york.ac.uk/prospero/, identifier: PROSPERO CRD42021233145.Copyright © 2023 Wang, Liu, Li, Li, Li, Hui, Li, Zhou and Yi.