The paper proposes an integrated system to automatically assess motor function after neurological injury. A portable motion capture system was developed in order to obtain all the relevant three dimensional kinematics of the upper limb movement. These kinematics were analyzed by means of a decision tree classifier which features where inferred from the Functional Ability Score (FAS) of the Wolf Motor Function Test (WMFT). In addition, the system is able to correctly quantify the performance time of each selected task of the WMFT. In terms of the FAS the system and the clinician show coherent results for 3 out of 5 patients in the first task tested and 4 out of 5 for the second task tested. Regarding performance time, the mean error between the system and the clinician was of 0.216 s for the 25 trials performed (5 patients, 5 tasks each). These results represent an important proof of concept towards a system capable of precisely evaluate upper limb motor function after neurological injury.