With the increasing penetration level of wind power and other renewable energy resources，the demands for peak- regulating ancillary services are rising rapidly. Currently，some problems about the peak-regulating auxiliary services exist in six major power grid regions such as the single standard of compensation for deep peak-regulating of the thermal power units， small and medium capacity units with weaker peak-regulating capacity can barely obtain compensation. The unfairness during the existing peak-regulating auxiliary service compensation mechanisms leads to the consequence that a large number of thermal power units lack the incentives of providing peak-regulating auxiliary services. K-means cluster analysis is used to classify the units based on the unit capacity and the highest peaking rate indicators，and then the boundary line between basic peak- regulating and deep peak-regulating across various units is identified. Moreover，a new auxiliary service compensation model is put forward to encourage more units providing peak-regulating auxiliary services during low-load period，thus increasing stability of the electricity grid as well as promoting more wind power integration into the grid. The case studies of63units dispatched by the Northeast Power Grid are given to verify the effectiveness of the improved model on encouraging more thermal power units to participate in the provision of peak-regulating auxiliary services.