This study investigated the diversity, density, and distribution of cattle-infesting flies in smallholder systems in the Malang region, East Java, Indonesia, where high fly burdens compromise animal health and productivity, but spatial evidence is limited. In June and December 2024, we surveyed 400 cattle across 13 sub-districts. Flies were identified morphologically, supported by a YOLOv8 computer vision workflow. Relative abundance, Shannon diversity (H′), body-site predilection, and associations with location, breed, body condition score (BCS), and housing/hygiene were analyzed, and spatial patterns were mapped. A total of 6,237 flies, representing nine distinct species, were recorded. Haematobia irritans exigua dominated the diversity (n = 2,180), followed by Musca domestica (1,455), Stomoxys spp. (S. bengalensis 1,083; S. sitiens 600; S. calcitrans 523), M. autumnalis (370), Morellia simplex (13), M. crassirostris (9), and Tabanus rubidus (4). The highest diversity was found in Malang Regency (H′ = 1.55). The highest relative abundance of H. irritans exigua was observed in Malang City (50.3%) and Malang Regency (44.7%), while M. domestica predominated in Batu City (43.8%). Haematophagous flies concentrated on the lower and upper body regions. Higher burdens were associated with black or brown solid coat color, fat cattle with high body condition scores (BCS 7–9), and the Simmental breed. H. irritans exigua predominated in individually housed cattle under poor hygiene conditions. The YOLOv8-based identification system demonstrated high performance (mAP 99.5%, precision 74.4%, recall 94.6%) in supporting microscopic identification. Further improvements, however, require additional model training with larger datasets.