1) to evaluate the efficacy of medical cannabinoids (MC) by appraising the quality of evidence from clinical studies 2) to explore the factors hampering the MC use in clinical practice of Parkinson's disease (PD). We performed a systematic review through various databases. The quality of 14 studies was assessed by Cochrane risk bias (5 randomized controlled trials- RCT) and Newcastle-Ottawa scale (9 uncontrolled studies). The positive effects on motor (5 studies) and non-motor symptoms (4 studies) described in uncontrolled studies have not been confirmed by the few and small RCTs. Only one RCT found a reduction of levodopa-induced dyskinesias, another a reduction in anxiety and tremor amplitude in an anxiogenic situation, while the remaining three without effect on motor/non-motor symptoms. Physical and psychological symptoms are among the most common side effects. There is insufficient evidence to reform international legislation regarding cannabis use in PD practice. •There is insufficient evidence to recommend the safe use of cannabis for motor symptoms in PD.•MC should not be encouraged, except for a trial or selected patients with refractory non-motor symptoms (such as pain).•Well- designed RCTs are needed to establish evidence, which will inform the so far inconsistent MC legislation in PD.