Using spectral, wavelet, multifractal, and recurrence analyses we examined the features of involuntary shaking (tremor) arising during the performance of a given motor task. The task was to maintain the efforts of fingers under isometric conditions by a healthy subject, a patient with primary bilateral parkinsonism, and a patient with a syndrome of the essential tremor. The physiological tremor was characterized by the minimal amplitude, the broad power spectrum, the minimal energy of the wavelet spectrum, the maximal degree of multifractality, the minimal degree of determinism, and the maximal entropy of recurrence time density. In the case of the essential tremor we observed the significant enhancement of the wavelet spectrum energy and the decrease of the oscillation complexity. It was evident as the appearance of clear peaks in the power spectra, the fall in the multifractality degree, the emergence of a quasiperiodic structure in the recurrence diagrams, the grow of determinism and the decrease of the entropy of recurrence time density. All these trends were enhanced for the parkinsonian tremor. The suggested characteristics enable us to estimate the degree of deviation of the motor function from the healthy one.