In this paper, we bring out modern methods that are potential in analysing differences in the dynamic surface electromyographic (EMG) and acceleration measurements between patients with Parkinson's disease (PD) and healthy persons. These methods are the correlation dimension of EMG, the recurrence rate of EMG, the power of acceleration and the sample entropy of acceleration. In this study, these methods were used to extract features from surface EMG and acceleration recordings measured during elbow flexion and extension movements. The extracted features were used to form high-dimensional feature vectors and the dimensionality of these vectors was then reduced by using the principal component approach. Finally, the feature vectors were discriminated between subjects by using the principal components. The discrimination power of the presented approach was tested with EMG and acceleration data measured from 46 patients with PD (on-medication) and 59 healthy controls. Discrimination results showed that the present method was able to discriminate dynamic EMG and acceleration recordings between patients with PD and healthy controls. Therefore, dynamic surface EMG and acceleration measurements may have potential in the objective and quantitative assessment and diagnosis of PD.