Abstract: The reasonable configuration of the discharge characteristic function of turbine governing valves is essential to linear power control for thermal power units. Then the discharge characteristic function of turbine governing valves shall be re-optimized when there exists any undesired nonlinearity of the discharge characteristics which may degrade the control performance. Currently, the optimization of the discharge characteristic function depends on the valve discharge characteristic experiment. In this paper, based on the test study of a 362.5 MW unit, a method of online valve discharge characteristics monitoring and optimization is proposed, which can observe the valve discharge characteristics and optimize the discharge characteristics function by using the historical DCS data. Furthermore, by virtue of this method the valve discharge characteristics function can be rectified in advance according to the changes of the valve discharge characteristics to avoid the performance degradation of primary frequency regulation and the power oscillations caused by the nonlinearity of the valve discharge characteristics.