Information Fusion plays important role in Prognostics and Health Management, where data and informations from different sources need to be combined, analyzed and finally used or presented for proper maintenance decisions. The objective of this paper is to outline and analyzed the relation between Information Fusion process and data/information processing in Prognostics Condition Based Maintenance. The Data Fusion Information Group Model (DFIGM) is presented as well as distinction between two levels of information fusion: (1) low-level information fusion (LLIF), which addresses the signal processing, object state estimation and characterization, and (2) high-level information fusion (HLIF) focused on control and situational understanding. All this processes are aligned with condition based maintenance processes from data acquisition and processing, through diagnostics, prognostics, up to health management. Presented work is one of the first steps in the research project toward improvements in Condition Based Maintenance with focus on its implementation in the manufacturing industry.