How to fully leverage the value of big data, achieve precise response and satisfy public service needs has become an important research topic. This paper based on the DIKW model and from the perspective of data flow, a public service demand management model based on big data of social appeals  is constructed. This model divides the entire demand management process into three consecutive stages: demand identification, demand analysis, and demand decision-making. Through structured description, aggregation analysis, and comprehensive evaluation of social appeals big data, precise management of public service demand is achieved. On this basis, collect data from local leaders' message boards on People's Daily Online for empirical analysis. It is shown that the proposed model and method are feasible and can comprehensively characterize public service demand, accurately discover identity category, regional distribution, emotional polarity and other characteristics of demand subjects, and provide a basis for demand decision-making.