This paper presents our research in real-time adaptive signal processing of physiological hand tremor with a DSP controller. A DSP controller can be applied in this application because of its high-performance in the signal processing and its capability in the realization of a single-chip embedded system for both real-time signal processing and control applications. The potential applications of the system include the long-term on-site monitoring of physiological tremor and the suppression of physiological tremor in a hand-held tool. Because of the limited memory and the operation speed provided by a DSP controller, recursive least-squares (RLS) adaptive algorithm is investigated in the modeling and on-line prediction of physiological hand tremor. The effectiveness of the algorithm is illustrated with the simulation and experimental results.