Aiming at the problems of strong data redundancy and inconsistent transmission protocol in digital manufacturing shop， an edge computing oriented real-time data fusion method is proposed. Firstly， around the idea of edge computing， the Internet of things based real-time workshop data acquisition architecture is established. Secondly， drawing on evidence theory and based on consistency detection and normalized data， the mutual support degree is proposed for evidence transformation， and the improved evidence combination rule is introduced to obtain fusion data， so as to reduce the amount of data uploaded by the device and reduce the transmission pressure of the transmission network. Then， based on the OPC UA information model， the information hierarchy model of the configuration required for protocol transformation is constructed， and the protocol transformation is completed in the form of visual configuration， providing a unified transmission interface for workshop data. Finally， the bus control console in a certain aerospace manufacturing workshop is used to verify the effectiveness of the proposed method.