The objective of this study is to assess the efficacy and safety of herbal medicines (HMs), as a monotherapy or adjunct therapy, compared to placebo or conventional approaches in the treatment of idiopathic Parkinson's disease (PD). We conducted a systematic review of randomized controlled trials from both conventional and alternative medicine sources. Outcome measures were overall improvement, quality of life, reduction of levodopa dose, and adverse events. Nine studies were included, each testing a different HM. Six of the trials had limited internal validity due to major flaws in design, including the lack of proper randomization; insufficient blinding; unclear inclusive criteria in terms of diagnostic criteria, baseline staging, and duration of disease; lack of proper sample size calculation; and insufficient data analysis. Imbalances in gender and ethnicity among the patients in the included trials were observed. No major adverse events emerged, and no specific pattern was detected from the trials describing such data. In addition to major methodological defects, heterogeneity in (1) HM tested, (2) control treatment, and (3) outcome measure hindered in-depth data analysis and synthesis. Current evidence is insufficient to evaluate the efficacy and safety of various HMs. Further studies with improved trial design and reporting, with assessment on cost-effectiveness, quality of life, and qualitative data are warranted.