At present, the number of coronavirus disease 2019(COVID-19) cases worldwide is increasing, the spatio temporal spread of the epidemic becomes more and more complicated. The traditional researches on the transmission process is mainly focus on transmission trends of infectious diseases at the macro level. It is impossible to analyze the transmission relationship between specific cases at the individual level, to accurately locate the transmission paths of the epidemic, and it is difficult to support the precise prevention of infectious diseases. So, it is an urgent need to study the transmission process of infectious diseases on both of the semantic and spatio-temporal features. Based on the analysis of COVID-19 epidemic cases data, we construct the COVID-19 cases activity knowledge graph, which adapts to various description methods. Then, we design the ontology rules to complete the construction of the pattern layer. The epidemiological survey data has been analyzed to recognize the event entities, and to complete the construction of data layer. Finally, through the graph database and the B/S pattern to build a prototype system for experimental verification. The results show that it is effective and feasible to analyze the transmission process, infection relationship, key nodes and activity trajectory through the COVID-19 cases activity knowledge graph.