User activity reconstruction is a technique used in digital forensic investigation. Using this technique, digital forensic investigators extract a list of user activities from digital artifacts confiscated at the crime scene. Based on the list, explicit knowledge about the crime, such as motive, method, time, and place, can be deduced. Until now, activity reconstruction has been conducted by manual analysis. This means that the domain of the reconstructed activities is limited to the personal knowledge of the investigators, so the result exhibits low accuracy due to human errors , and the process requires an excessive amount of time. To solve these problems, this paper proposes a digital forensic framework SigDiff for automated user activity reconstruction. This framework uses a signature-based approach. It comprises an activity signature generation module, signature database, digital artifact collection module, and activity reconstruction module. Using SigDiff, the process of user activity reconstruction can be performed accurately with a high retrieval rate and in a reduced time span.