Digital forensics is a prime professional field for law enforcement organizations. This is also a major active research topic in cybersecurity. Although traffic and content analysis are leading tasks in this field, most Internet traffic is now encrypted, making traditional content analysis impossible. Furthermore, instant messaging (IM) applications have become increasingly popular for communication between individuals and groups. However, IM conversations can be used for illicit activities such as planning criminal activities or exchanging sensitive information. In such cases, law enforcement agencies may need to perform VoIP forensics to identify suspects involved in conversations. This study proposes a network forensic approach (NFA) for correlating IM calls to identify the IP addresses of suspects. The approach involves capturing and analyzing IM call data, correlating the data with other network traffic, and using the correlation to identify suspects' IP addresses. The proposed approach was tested on real-world IM call data and yielded promising results. The network forensics approach for VoIP is superior to other approaches that need physical access to the end-users' devices, making NFA suitable for early crime detection and in situations where the devices may have been destroyed or burnt. The proposed method attained a success rate of 92.5% in identifying voice IM calls and providing information about the participants involved in the calls.