Abstract: Classical scholarship,grounded in studies of Western European nation-states,has developed “state capacity” into a central concept in the social sciences and widely applied it to the analysis of developing countries. However,this body of cross-national research,primarily concerned with the origins of state capacity,often overlooks the distinctive logic of capacity formation in the Third World. As a result,it struggles to explain why,in the absence of significant changes in structural conditions,the same country may exhibit markedly different governance outcomes across historical periods or policy domains. Taking this puzzle as its point of departure,this article adopts a process-mechanism approach to trace causal linkages among concrete actions in governance. It conceptualizes state capacity as a form of governing potential and emphasizes how,under conditions of multiple concurrent tasks and structural tensions,the state transforms potential into effective capacity through goal adjustment,resource concentration,and experiential learning. The level of performance in this transformation depends on whether state actions at different stages can maintain a dynamic equilibrium among political coordination,administrative implementation,and societal response. Such equilibrium arises from the interaction and recombination of multiple mechanisms. In this way,the article moves beyond explanations that attribute variations in state capacity to static structural factors such as war or institutional design,and instead treats it as the outcome of a dynamic,interactive process. This shift not only enhances the explanatory power for governance variation in developing countries but also provides a testable analytical framework for understanding China's experience of state governance modernization.