The study of the Reaching movements, since their significant importance for independence in daily living activities, is increasingly proposed in rehabilitation programs of several disorders of upper limbs. Robotic rehabilitation programs represent a valid approach to this issue, as they are well-matched to realize an rigorous, task-oriented motor training as part of an combined rehabilitation program, counting also non-robotic approaches. However, there is a lack of information showing progresses on parameters related to activities of daily living (ADL), and some experimental studies indicates that to date robotic training fails to transfer the obtained improvements on the functional level. This may be since most of the existing robotic devices are based on the repetition of stereotypical movements rather than of applying more appropriated principles of motor learning. This work proposes a new quantitative kinematic index to assess the kinematic quality of upper arm reaching movements based from the mathematical minimum jerk theory. The results showed as the indexes are close to the ideal ones in according with the mathematica model, making them suitable in a clinical contest for the valuation of the ADL performance during the rehabilitation of upper limb reaching movements from several diseases.