Objective To evaluate the effectiveness of different types of physiotherapy interventions in people with Parkinson’s disease (PD). Design Systematic review and meta-analysis of randomized controlled trials (RCTs). Methods Five databases (PubMed, Embase, Cochrane Library, CINAHL and Web of Science Core Collection) were searched for relevant RCTs published from database inception to July 14, 2022. Reviewers independently screened the literature, extracted data, and assessed the literature quality according to the Cochrane Collaboration Risk of Bias Tool and PEDro Scale. This meta-analysis was conducted using RevMan 5.4.1 and reported in compliance with the PRISMA statement. Results Forty-two RCTs with 2,530 participants were included. Across all types of physiotherapy, strength training, mind-body exercise, aerobic exercise, and non-invasive brain stimulation (NiBS) were effective in improving motor symptoms as measured by the (Movement Disorders Society-) Unified PD Scale, whereas balance and gait training (BGT) and acupuncture were not. The pooled results showed that the change in mind-body exercise (MD = −5.36, 95% CI [−7.97 to −2.74], p < .01, I2 = 68%) and NiBS (MD = −4.59, 95% CI [−8.59 to −0.59], p = .02, I2 = 78%) reached clinical threshold, indicating clinically meaningful improvements. Considering the effectiveness of the interventions on motor symptoms, balance, gait and functional mobility, mind-body exercise was recommended the most. Conclusions Exercise appears to be a better form of physiotherapy than NiBS and acupuncture for improving motor function. Mind-body exercise showed beneficial effects on motor symptoms, balance, gait and functional mobility in people with PD, and is worthy of being promoted.