Uluguru Mountains is one of the Eastern Arc Mountain blocks that stretch from Taita Hills in Kenya to Udzungwa Mountains in south-central Tanzania. The Uluguru Mountains contain several forest fragments, which are recognized as important biodiversity hotspots. The population pressure and encroachment threatens their biodiversity values. This study analyses (1) species richness, diversity, floristic similarity and structure of trees species (DBH ≥ 10 cm) in the forest fragments in Uluguru, (2) Understory species composition and natural regeneration of the fragments (3) changes in species richness, diversity and tree density along the edge-interior gradient in the fragments and (4) indigenous tree use, use values and human population impacts on tree density and species richness in Uluguru forest fragments.  In the overstory layer (i.e. DBH ≥ 10 cm) examination of 900 individual trees revealed 101 species, 73 genera and 31 families. Fabaceae and Moraceae were the most speciose and important families in terms of familial importance value (FIV). The five species with the highest species importance value (IVI) were Ehretia amoena, Khaya anthotheca, Synsepalum cerasiferum, Sorindeia madagascariensis, Diospyros squarrosa, all accounting for 27 % of the total IVI. Of the total observed species, 31 % had up to two individuals while 15 % occurred only once in all the sample plots. The studied forests differed significantly in their species richness (26 - 93 species ha-1), tree density (85 - 390 stems ha-1), basal area (3 - 24 m2 ha-1) and diversity indices (Shannon-Wiener 2.50 - 4.02; Fishers 6.8 - 35.5; and Simpson 9.4 - 63.1). The forest fragment size revealed positive correlation with species richness (r = 0.92), trees density (r = 0.66) and basal area (r = 0.28). In general, the larger forests (Kilengwe and Kimboza) had higher species richness, which might be due to high environmental heterogeneity within the forests. However, the small fragments should be given priority in conservation to avoid loss of species. The dendogram revealed four clusters of forests with low similarity between them, the closer the forests were to one another, the more similar in floristic composition. The significant negative association between geographical distance and floristic similarity among forest pairs was confirmed by correlation test (r = -0.43, p < 0.001). Although, all studied forests showed good regeneration pattern, more attention on conservation should be paid to rare, threatened species and those with lower IVI.  In the understory layer, a total of 2119 seedlings, 1798 small saplings and 2585 large saplings representing 91, 102 and 104 species belonging to 28, 32 and 30 families respectively were recorded in the studied forests. Fabaceae and Moraceae were the most speciose families in the the entire understory compartments. Sorindeia madagascariensis, Scorodophloeus fischeri, Diospyros squarrosa and Ehretia amoena appeared in the top ten species with the highest IVI in all the three compartments. The forests varied significantly in terms of their species richness, diversity indices and structural characteristics in all the three compartments. The overall frequency distribution of the species from the three forest compartments revealed high species richness in the two lower frequency classes (i.e. < 20 % and 20-40 %) in all the forests indicating a high degree of floristic heterogeneity and the presence of rare species within the forests. Also, this study revealed the understory layer being richer than overstory in all the forest fragments. The overall mean species richness for understory was 59 species ha-1 while overstory had 47 species ha-1. The Srensen coefficients of similarity between understory and overstory layers ranged from 0.53 to 0.77, and were higher within fragments than between them. Though, all forests displayed good regeneration, some species like Dalbergia melanoxylon, Pterocarpus angolensis, Pandanus rabaiensis, Millettia sacleuxii, Pouteria altissima, Allanblackia stuhlmannii revealed poor regeneration, hence require due attention.  The variation of species richness, diversity and density of tree community along edge-interior gradient were studied in all the surveyed forests. A total of 198 individual ha-1 representing mature trees, 566 individuals ha-1 large saplings, 6309 individual ha-1 small saplings and 46469 individuals ha-1 seedlings were encountered in all the seven studied forests. In general, the mean species richness, diversity and tree density in the interior plots of the overstory, large saplings and small sapling layers were significantly higher than edge and intermediate plots that did not differ appreciably. The regression slopes of the interior plots in overstory and large sapling were significantly higher than edge and intermediate plots. For small saplings and seedlings, interior and intermediate did not differ appreciably in their regression slopes but were considerably higher than edge. The comparison of species richness, diversity and density within each categorical distance showed significant variation in all the forest layers. The relationship between tree size classes and tree density from edge and interior plots in all the forests revealed exponential decay patterns. This implies that some trees die due to competitive advantage from crowding and suppression as the stand approach a limiting number of trees of a given size that may coexist within a given area. The findings of this section concluded that the interior of the forests and the edges/intermediate were contrasting habitats. Moreover, long-term study is needed to investigate how microclimatic conditions affect species richness, composition and density within the studied forest regions.  Assessment of tree uses and use values was carried out using structured questionnaires. A total of 42 species belonging to 38 genera representing 19 families were listed by respondents as being useful for their livelihoods. Of these species, 88 % were listed to serve more than one function. Milicia excelsa, Albizia gummifera and Annona senegalensis had higher total use values among the identified species whereas Sterculia quinqueloba had the lowest use value. 64 % of listed species are used for both firewood and charcoal making. Species used for timber and medicinal purposes accounted for 45 % and 40 % of the identified species respectively. Species used for building poles and making domestic utensils accounted for 55 % each while those used for carving works and fodder, accounted 40 % and 29 % of all listed species respectively. Moreover, this study revealed that the human population surrounding the forests was negatively correlated with forest size (r = -0.90), species richness (r = -0.79) and tree density (r = -0.76). The results suggest that the increase of human population will continue putting pressure to forest resources because of increased demand for more land for agriculture, charcoal and firewood for fuel, timber and poles for building puroposes and medicines for their health. Moreover, the study suggests that indigenous knowledge is vital ingredient in the on-going effort to reverse the trend of environmental degradation in the Uluguru forests by integrating it into community based management plans. Also, there is a need to control human population in the villages surrounding the forests for the purpose of reducing pressure towards forests resources.  In general, the study revealed that habitat fragmentation is associated with drastic changes in the species composition and structure of the forests. If fragmentation process continues, the ability of forest remnants to sustain their original biodiversity and ecological processes will be considerably reduced. Thus, protection of these fragments needs to be prioritized.