In response to the increasing requirements of the power grid for the regulation capability of generating units，it is necessary to explore effective ways to enhance Automatic Generation Control(AGC)performance within the framework of smartpower plants. Taking a 1 000 MW ultra-supercritical coal-fired unit of a thermal power enterprise as the research object，this study delves into the deficiencies of the AGC system in terms of response speed，regulation accuracy，and adaptability to complexoperating conditions. It systematically analyzes constraining factors such as coordinated control strategies，equipment response characteristics， and control parameter settings，and proposes an improvement scheme based on feature modeling and comprehensive optimization. The results indicate that through targeted technical modifications and control logic redesign，the regulation performance of the unit can be significantly enhanced，thereby strengthening its capability to participate in peak shavingancillary services for the power grid and providing practical references for the intelligent development of the power system.