Parkinson’s disease is primarily marked by mitochondrial dysfunction and metabolic abnormalities. We recently reported that the combined metabolic activators improved the immunohistochemical parameters and behavioural functions in Parkinson’s disease and Alzheimer’s disease animal models and the cognitive functions in Alzheimer’s disease patients. These metabolic activators serve as the precursors of nicotinamide adenine dinucleotide and glutathione, and they can be used to activate mitochondrial metabolism and eventually treat mitochondrial dysfunction. Here, we designed a randomized, double-blinded, placebo-controlled phase II study in Parkinson’s disease patients with 84 days combined metabolic activator administration. A single dose of combined metabolic activator contains L-serine (12.35 g), N-acetyl-L-cysteine (2.55 g), nicotinamide riboside (1 g) and L-carnitine tartrate (3.73 g). Patients were administered either one dose of combined metabolic activator or a placebo daily for the initial 28 days, followed by twice-daily dosing for the next 56 days. The main goal of the study was to evaluate the clinical impact on motor functions using the Unified Parkinson’s Disease Rating Scale and to determine the safety and tolerability of combined metabolic activator. A secondary objective was to assess cognitive functions utilizing the Montreal Cognitive Assessment and to analyse brain activity through functional MRI. We also performed comprehensive plasma metabolomics and proteomics analysis for detailed characterization of Parkinson’s disease patients who participated in the study. Although no improvement in motor functions was observed, cognitive function was shown to be significantly improved (P < 0.0000) in Parkinson’s disease patients treated with the combined metabolic activator group over 84 days, whereas no such improvement was noted in the placebo group (P > 0.05). Moreover, a significant reduction (P = 0.001) in Montreal Cognitive Assessment scores was observed in the combined metabolic activator group, with no decline (P > 0.05) in the placebo group among severe Parkinson’s disease patients with lower baseline Montreal Cognitive Assessment scores. We showed that improvement in cognition was associated with critical brain network alterations based on functional MRI analysis, especially relevant to areas with cognitive functions in the brain. Finally, through a comprehensive multi-omics analysis, we elucidated the molecular mechanisms underlying cognitive improvements observed in Parkinson’s disease patients. Our results show that combined metabolic activator administration leads to enhanced cognitive function and improved metabolic health in Parkinson’s disease patients as recently shown in Alzheimer’s disease patients.