Ground-dwelling birds play vital roles in forest ecosystems, such as seed dispersal and pest control, but are vulnerable to habitat fragmentation and hard to detect using traditional survey methods. Camera traps offer a non-invasive, effective alternative for monitoring these elusive species, especially in difficult terrain and over extended periods. Camera trapping can enhance data collection on species behaviour, distribution and diversity, especially on rare, elusive species. The Albertine Rift Valley, a biodiversity hotspot, is home to many endemic bird species, yet their ecology remains poorly understood. A study in Rwanda's Gishwati-Mukura National Park used camera traps to assess bird diversity and compare two forest fragments. The study used 64 camera traps arranged in 4 × 4 grids across Gishwati and Mukura Forests from May 2017 to May 2018. Images were analysed for ground-dwelling birds using defined photographic events. Species richness was assessed through rarefied species accumulation curves and Jacknife-2 estimators. Diversity was measured using Shannon, Simpson, and evenness indices, which were subsequently compared between forest fragments. Bird community composition was examined using Detrended Correspondence Analysis. Eighteen bird species were recorded, with similar richness and diversity in either forest, though Mukura showed greater species heterogeneity, that is, greater variation in species composition. Detection rates were limited due to insufficient sampling effort, suggesting true species richness may well exceed 20 species. Some rare and endemic species were documented, highlighting camera trapping to be a valuable supplement to traditional survey methods. Despite habitat disturbance, both forests maintain healthy bird communities. However, more extensive sampling is needed to improve precision and to better understand community structure and species turnover between forest fragments.