Mild cognitive impairment (MCI) is a very common non-motor feature of Parkinson’s disease (PD) and the non-amnestic single-domain is the most frequent subtype. Transcranial random noise stimulation (tRNS) is a non-invasive technique, which is capable of enhancing cortical excitability. As the main contributor to voluntary movement control, the primary motor cortex (M1) has been recently reported to be involved in higher cognitive functioning. The aim of this study is to evaluate the effects of tRNS applied over M1 in PD-MCI patients in cognitive and motor tasks. Ten PD-MCI patients, diagnosed according to the Movement Disorder Society, Level II criteria for MCI, underwent active (real) and placebo (sham) tRNS single sessions, at least 1 week apart. Patients underwent cognitive (Digit Span Forward and Backward, Digit Symbol, Visual Search, Letter Fluency, Stroop Test) and motor assessments (Unified Parkinson’s Disease Rating Scale [UPDRS-ME], specific timed trials for bradykinesia, 10-m walk and Timed up and go tests) before and after each session. A significant improvement in motor ability (UPDRS-ME and lateralized scores, ps from 0.049 to 0.003) was observed after real versus sham tRNS. On the contrary, no significant differences were found in other motor tasks and cognitive assessment both after real and sham stimulations. These results confirm that tRNS is a safe and effective tool for improving motor functioning in PD-MCI. Future studies using a multisession tRNS applied over multitargeted brain areas (i.e., dorsolateral prefrontal cortex and M1) are required to clarify the role of tRNS regarding rehabilitative intervention in PD.