Although Levodopa-induced dyskinesias (LID) are one of the most compromising complications of dopaminergic treatment in Parkinson's disease (PD), there is no widely accepted assessment tool available that evaluates LID quantitatively. This is of relevance as objective assessment may help to facilitate proof-of-concept studies with novel treatments and thus eventually contribute to better patient care. PD patients were asked to perform a grip-lift task as well as tapping tasks assessed with the "Q-Motor" system. PD patients were separated into three groups according to their modified abnormal involuntary movement scale (M-AIMS)-score: PD patients without dyskinesias (PD(LID-) n = 17), with slight dyskinesias (PD(LID+) n = 15) and with severe dyskinesias (PD(LID++) n = 15). An explorative analysis to identify measures detecting LID was performed with 5 PD(LID-) and 5 PD(LID++) patients; these measures were then used in the remaining patients to assess the accuracy of the system to differentiate LID. The measures "Orientation-Index" and "Position-Index" of the grip-lift task differed significantly between the explorative cohorts. Using these two parameters for the differentiation of the remaining cohorts, the area under the ROC curve (AUC) yielded 0.809 for the differentiation of PD(LID-) vs. PD(LID++), 0.852 for the differentiation of PD(LID-) vs. PD(LID+) patients, and 0.830 for the differentiation of PD(LID+) and PD(LID++). The "Orientation-Index" and "Position-Index" of the Q-Motor assessment are sensitive, easy to apply and non-invasive measures for the objective assessment of manifestation and severity of LID.