Public health emergencies in the big data environment have the characteristics of massive information sources and multiple types of information. As a new method of constructing a multi-dimensional and cross-media information aggregation system, information aggregation provides a new idea for the orderly arrangement and utilization of the information. By analyzing and summarizing the risk research and judgment process of public health emergencies and the decision-making information needs and sources of event participants, a semantic aggregation model of public health emergencies information based on semantic relationships is constructed. The overall design of the aggregation mode is divided into data, the processing layer, the semantic aggregation layer and the service application layer are three parts, namely from data processing, information aggregation and resource application, to realize the methodological exploration of how to construct the semantic aggregation mode of decision-making information for public health emergencies. At the same time, taking the SARS virus infection incident in early 2003 as an example, the event data and information in the decision-making information were aggregated at the semantic level to verify decision-making information for emergencies. Related research provides new ideas and development directions.