BACKGROUND: The treatment for Parkinson's disease (PD) consumes a lot of manpower and financial resources. Non-pharmacological alternative motor-/sensory-based interventions are optimized for the rehabilitation of PD patients. Mindfulness-based therapy shows ideal efficacy, but the diversity of the therapy brings difficulties to the selection of clinicians and patients. METHODS: Network meta-analysis in the Bayesian framework was used to evaluate the efficacy of non-pharmacological alternative motor-/sensory-based interventions in improving motor and non-motor symptoms in PD patients. RESULTS: A total of 58 studies (2,227 patients) were included. Compared with the non-intervention group, qigong was associated with improved outcomes in the Timed Up and Go (TUG) test (mean difference (MD) -5.54, 95% confidence interval (CI) -8.28 to -2.77), and UPDRS-I (MD -15.50, 95% CI -19.93 to -7.63). Differences between non-pharmacological alternative motor-/sensory-based interventions were not significant for PDQ-39, UPDRS-I, or UPDRS-II; however, qigong was superior to dance (MD -3.91, 95% CI -6.90 to -0.95), Tai Chi (MD -3.54, 95% CI -6.53 to -0.69), acupuncture (MD -6.75, 95% CI -10.86 to -2.70), music (MD -3.91, 95% CI -7.49 to -0.48), and exercise (MD -3.91, 95% CI -6.49 to -1.33) in the TUG test. CONCLUSION: This network meta-analysis supports mindfulness-based therapy (e.g., qigong, yoga, and Tai Chi) as a preferred non-pharmacological alternative motor-/sensory-based intervention for PD rehabilitation. SYSTEMATIC REVIEW REGISTRATION: https://inplasy.com/inplasy-2022-10-0109/, INPLASY2022100109.