Abstract Data provenance is defined as the origin and process history of a derived data item, which is becoming increasingly important for scientific research work. Although provenance can be very valuable for applications, research on data provenance is on its beginning and there is still a series of work that needs to be done. In this paper, some categories, which are provenance granularity, data granularity, data status, provenance computing, semantics of provenance, provenance storage and applications, are proposed to classify and analyze present provenance techniques. In this work we introduce a provenance classification for different applications based on proposed categories, to help the researchers in this field to understand the existing problems and the current challenges and how to enable the current techniques for solving these challenges.