In order to fully utilize different types of time transfer links, it is necessary to realize the fusion application of time transfer link data under different sampling rates. This article proposes a data fusion method based on multi-resolution analysis. Firstly, we perform wavelet decomposition on the original data to achieve a unified resolution and initially eliminate high-frequency noise. Subsequently, Kalman filtering is employed across various resolutions. Ultimately, the Mallat fast reconstruction algorithm is utilized to derive the fusion result. The method is used to process time transfer data between the National Time Service Center (NTSC) of the Chinese Academy of Sciences and the Physikalisch-Technische Bundesanstalt (PTB) of Germany. The results show that the fusion algorithm can handle data problems caused by link anomalies or interruptions. Since the actual measurement results of the GPS (Global Positioning System) PPP (Precise Point Positioning solutions) link perform better than those of the TWSTFT (Two-Way Satellite Time and Frequency Transfer) link overall, the GPS PPP link measurement results are used to evaluate the gain of the fusion algorithm. Using Rapid Realization of Coordinated Universal Time (UTCr) as the reference, the accuracy gain of the data fusion result is about 1%, and the daily frequency stability gain is better than 20%. At the same time, the fusion algorithm can suppress the periodic noise of the TWSTFT link, effectively improving the stability and robustness of the link.