Background: In Parkinson's disease, reaching movements are conditioned by motor planning and execution deficiency. Recently, rehabilitation, aided by high technological devices, was employed for Parkinson's disease. Objective: We aimed to (1) investigate the changes in the upper limb motor performances in a sample of a patient with Parkinson's disease after a weightless training, with a passive exoskeleton, in an augmented-feedback environment; (2) highlight differences by motor parameters (performance, speed, and movement accuracy) and by type of movement (simple or complex); and (3) evaluate movement improvements by UPDRS II-III. Methods: Observational pilot study. Twenty right-handed patients with Parkinson's disease, Hohen and Yahr 2, Mini Mental State Examination ≥24 were evaluated. All patients underwent 5 day/week sessions for 4 weeks, 30 min for each arm; the training was performed with 12 exercises (single and multi-joints, horizontal and vertical movements). All the patients were assessed by UPDRS II-III and the evaluation tests provided by the device's software: a simple movement, the vertical capture, and a complex movement, the horizontal capture. For each test, we analyzed reached target percentage, movement execution time, and accuracy. Results: After training, a significant improvement of accuracy and speed for simple movement on the dominant arm, of reached targets and speed for complex movement on both sides were shown. UPDRS II and III improved significantly after training. Conclusions: In our study, a motor training aided by a high technological device improves motor parameters and highlights differences between the type of movement (simple or complex) and movement parameters (speed and accuracy) in a sample of patients with Parkinson's disease.