Bus bunching is a common phenomenon during bus operation. Previous studies on bus bunching mainly focus on the effect of single bus line. However, it is often the scene when several bus lines running in one corridor, especially in the downtown areas in China. Under this circumstance, different bus lines will interact with each other, and as a result, aggravating the bus bunching. The development of big data and data processing technology has brought new opportunities to the research of multi-line bus bunching. Based on the summary of relevant research, we did some essential processing and analysis on bus big data of Foshan City and chose the bus lines with obvious bunching phenomenon, high level of collinear and good data quality as analysis object in this paper. Real-time holding strategies for single bus line and two bus lines were established respectively. Through simulation, the bus operation status can be monitored real time and then it is judged and decided whether to hold the bus at the stop. The index calculation results prove the effectiveness of both strategies on bus bunching. A comparison between the two strategies shows that the holding strategy for single bus line makes more improvement than traditional strategies. But when it comes to aspects such as holding frequency, holding time, total running time and the influence on the other bus lines in the same corridor, the two-bus cooperative control strategy performs better. It demonstrates that it's meaningful to research into the influence of corridor service on bus bunching and take that into account to improve the traditional holding strategies. The strategies proposed in this paper have great significance on enriching the existing bus system analysis theory and method and improving the service level and intelligent management decision support system of public transit system. key words: bus big data, bus bunching, real-time holding strategy, multi-bus cooperative control

CICTP