Roof falls are common problems of underground coal mines which may have serious effects on workers in the form of injury, disability or fatality as well as other damages like production downtimes, interruptions in mining operations, equipment breakdowns, etc. Currently there is no system in place for continuous monitoring of roof health status of an underground Bord & Pillar coal mine to forecast roof fall events. This paper addresses a multi-hop wireless sensor network (MWSN) based solution for real time monitoring of roof health status in an underground Bord & Pillar coal mine. The MWSN is a unique system developed for monitoring roof health status of underground coal mines. Based on obtained data, mine management would be able to take appropriate decision on roof fall events and notify the workers about the roof health status from time to time. The paper presents the performance of a wireless sensor network used for assimilating roof vibration data from underground to a surface computer and analyzes roof fall events from the monitored data.