The extensive nature of ostrich farming production systems bears the continual risk of point introductions of avian influenza virus (AIV) from wild birds, but immune status, management, population density, and other causes of stress in ostriches are the ultimate determinants of the severity of the disease in this species. From January 2012 to December 2014, more than 70 incidents of AIV in ostriches were reported in South Africa. These included H5N2 and H7N1 low pathogenicity avian influenza (LPAI) in 2012, H7N7 LPAI in 2013, and H5N2 LPAI in 2014. To resolve the molecular epidemiology in South Africa, the entire South African viral repository from ostriches and wild birds from 1991 to 2013 (n = 42) was resequenced by next-generation sequencing technology to obtain complete genomes for comparison. The phylogenetic results were supplemented with serological data for ostriches from 2012 to 2014, and AIV-detection data from surveillance of 17 762 wild birds sampled over the same period. Phylogenetic evidence pointed to wild birds, e.g., African sacred ibis (Threskiornis aethiopicus), in the dissemination of H7N1 LPAI to ostriches in the Eastern and Western Cape provinces during 2012, in separate incidents that could not be epidemiologically linked. In contrast, the H7N7 LPAI outbreaks in 2013 that were restricted to the Western Cape Province appear to have originated from a single-point introduction from wild birds. Two H5N2 viruses detected in ostriches in 2012 were determined to be LPAI strains that were new introductions, epidemiologically unrelated to the 2011 highly pathogenic avian influenza (HPAI) outbreaks. Seventeen of 27 (63%) ostrich viruses contained the polymerase basic 2 (PB2) E627K marker, and 2 of the ostrich isolates that lacked E627K contained the compensatory Q591K mutation, whereas a third virus had a D701N mutation. Ostriches maintain a low upper- to midtracheal temperature as part of their adaptive physiology for desert survival, which may explain the selection in ratites for E627K or its compensatory mutations-markers that facilitate AIV replication at lower temperatures. An AIV prevalence of 5.6% in wild birds was recorded between 2012 and 2014, considerably higher than AIV prevalence for the southern African region of 2.5%-3.6% reported in the period 2007-2009. Serological prevalence of AI in ostriches was 3.7%, 3.6%, and 6.1% for 2012, 2013, and 2014, respectively. An annual seasonal dip in incidence was evident around March/April (late summer/autumn), with peaks around July/August (mid to late winter). H5, H6, H7, and unidentified serotypes were present at varying levels over the 3-yr period.