Buses are the most popular means for urban mass transportation. With the development of urbanization and motorization, the population of cities and the scale of regions have been on the rise, which has brought greater demands for public transportation and higher requirements for public traffic system. Bus arrival time, as one of the important parameters in urban intelligent public transport system service, is the basic information that public transport travelers are most concerned about. To meet the needs of the bus travelers to query the bus ar- rival time in real time, this paper explores the design of real-time arrival time query system of city buses. Based on the requirement analysis, the data model of the query system was established through the design of conceptual and logical structure. Then, the prediction of arrival time algorithm was put forward, which divided the route into the station sections and obtained the running time of each section from the historical records based on the data of bus arrival and departure collected in real time. The predicted running time was confirmed by clustering anal- ysis and discriminant analysis. Finally, the prediction of arrival time for users was obtained by simple calculation and the arrival time was displayed to the system users through the network.