Several nucleic acid-based assays have been developed for detecting Anaplasma marginale and Anaplasma centrale in vectors and hosts, making the choice of method to use in endemic areas difficult. We evaluated the ability of the reverse line blot (RLB) hybridisation assay, two nested polymerase chain reaction (nPCR) assays and a duplex real-time quantitative polymerase chain reaction (qPCR) assay to detect A. marginale and A. centrale infections in cattle (n = 66) in South Africa. The lowest detection limits for A. marginale plasmid DNA were 2500 copies by the RLB assay, 250 copies by the nPCR and qPCR assays and 2500, 250 and 25 copies of A. centrale plasmid DNA by the RLB, nPCR and qPCR assays respectively. The qPCR assay detected more A. marginale- and A. centrale-positive samples than the other assays, either as single or mixed infections. Although the results of the qPCR and nPCR tests were in agreement for the majority (38) of A. marginale-positive samples, 13 samples tested negative for A. marginale using nPCR but positive using qPCR. To explain this discrepancy, the target sequence region of the nPCR assay was evaluated by cloning and sequencing the msp1β gene from selected field samples. The results indicated sequence variation in the internal forward primer (AM100) area amongst the South African A. marginale msp1β sequences, resulting in false negatives. We propose the use of the duplex qPCR assay in future studies as it is more sensitive and offers the benefits of quantification and multiplex detection of both Anaplasma spp.