With the significant progress of information technology and artificial intelligence， digital twin has attracted more and more attentions because it provides a way of spacecraft design scheme evaluation， optimization and reliability analysis. In this paper， a multi-granularity digital twin model is established， and a multi-granularity data comparison and model updating method is proposed. By dividing the model granularity， the digital twin model is iteratively updated according to the granularity from coarse to fine. Through the real onboard orbit telemetry data of BeiDou M5， the data comparison and model updating of BeiDou M5 digital twin model are realized. The results show that the layered comparison method used in this paper can be applied to the digital twin model of BeiDou M5 satellite and achieved good results.