Objective: To identify the motor and non-motor factors that are associated with health related quality of life (HR-QOL) in a subgroup of Parkinson's disease (PD) patients with Levodopa therapy in early clinical stages. Methods: 391 Levodopa exposed patients were evaluated during the baseline assessment of a clinical trial in China. HR-QOL was measured by the Short Form 36 (SF-36). Motor and non-motor variables were determined during a structured interview and by clinical examination by movement disorder specialists. Multiple regression analyses were used to determine which variables were associated with low levels of HR-QOL. Results: Even if excluding non-motor variables from the regression model, motor factors, particularly motor deficits (measured by motor score of UPDRS), rigidity (measured by item 22 of UPDRS), and disease severity (measured by Hoehn&Yahr scale), explained only 18.9% of the variance of total SF-36 score. Whereas, when non-motor variables were included in the model, especially depression (measured by CES-D), sleep disturbances (measured by PSQ-1). and fatigue (measured by FSS), 61.7% of the variance of SF-36 score Could be explained. Two motor variables, UPDRS motor score and Hoehn&Yahr score, were also contributed to the model, however, the 95% confidence intervals (CIs) of these two motor factors were wide and included the null value (CIs -0.282, 0.019 for UPDRS motor score, and CIs -4.043, 0.856 for Hoehn&Yahr score). Neither, did higher daily levodopa dose contribute significantly to both models predicting SF-36 score. Conclusions: In our sample patients with levodopa therapy, motor disability and severity of parkinsonism contributed to a lesser extent to patients' self-report distress, within the first 5 years of disease onset. The clinical factors that showed the highest predictive value for worsen HR-QOL were non-motor symptoms, such as depression, sleep disorders, and fatigue. Great effort should be made to recognize and treat those conditions, thus improving all aspects of PD and giving these patients as a good HR-QOL as possible. (C) 2009 Elsevier Ltd. All rights reserved.