BACKGROUND Malaria transmission in Ethiopia is unstable and variable, caused by both Plasmodium falciparum and Plasmodium vivax. The Federal Ministry of Health (FMoH) is scaling up parasitological diagnosis of malaria at all levels of the health system; at peripheral health facilities this will be through use of rapid diagnostic tests (RDTs). The present study compared three RDT products to provide the FMoH with evidence to guide appropriate product selection. METHODS Performance of three multi-species (pf-HRP2/pan-pLDH and pf-HRP2/aldolase) RDTs (CareStart, ParaScreen and ICT Combo) was compared with 'gold standard' microscopy at three health centres in Jimma zone, Oromia Regional State. Ease of RDT use by health extension workers was assessed at community health posts. RDT heat stability was tested in a controlled laboratory setting according to WHO procedures. RESULTS A total of 2,383 patients with suspected malaria were enrolled between May and July 2009, 23.2% of whom were found to be infected with Plasmodium parasites by microscopy. All three RDTs were equally sensitive in detecting P. falciparum or mixed infection: 85.6% (95% confidence interval 81.2-89.4). RDT specificity was similar for detection of P. falciparum or mixed infection at around 92%. For detecting P. vivax infection, all three RDTs had similar sensitivity in the range of 82.5 to 85.0%. CareStart had higher specificity in detecting P. vivax (97.2%) than both ParaScreen and ICT Combo (p < 0.001 and p = 0.05, respectively). Health extension workers preferred CareStart and ParaScreen to ICT Combo due to the clear labelling of bands on the cassette, while the 'lab in a pack' style of CareStart was the preferred design. ParaScreen and CareStart passed all heat stability testing, while ICT Combo did not perform as well. CONCLUSIONS CareStart appeared to be the most appropriate option for use at health posts in Ethiopia, considering the combination of quantitative performance, ease of use and heat stability. When new products become available, the choice of multi-species RDT for Ethiopia should be regularly re-evaluated, as it would be desirable to identify a test with higher sensitivity than the ones evaluated here.