Abstract Background The Belgian Reference Laboratory for Plasmodium offers a free-of-charge reference testing of malaria-positive or doubtful samples to clinical laboratories. Methods The final malaria diagnosis from the Reference Laboratory (microscopy, rapid diagnostic tests (RDTs) and Plasmodium species-specific PCR) were compared with the final diagnosis from peripheral Belgian laboratories. The Reference Laboratory reports were analysed for all samples submitted between 2013 and 2017. Criteria assessed included the diagnosis of malaria, Plasmodium species identification including mixed infections, and in case of Plasmodium falciparum, the parasite density and the presence of sexual and asexual stages. Results A total of 947 non-duplicate samples were included. Reference testing confirmed 96.3% (893/927) and 90.0% (18/20) samples submitted as positive and negative, respectively, the two missed diagnoses were samples with Plasmodium ovale and Plasmodium malariae. Submitting laboratories had correctly identified P. falciparum in 95.1% (508/534) samples with P. falciparum single infection. They had correctly diagnosed the species in 62.9% (95/151) single non-falciparum samples and had reported ‘non-falciparum’ in another 26 (17.2%) samples; most errors occurred among P. malariae (n = 8/21, 38.1%) and P. ovale (n = 14/51, 27.5%). Only one of the 21 mixed Plasmodium species infections had been diagnosed as such by the submitting laboratories; in three of them, P. falciparum had been overlooked. Taken single and mixed infections together, P. falciparum was diagnosed in 98.6% (546/554) samples. Among 471 single P. falciparum samples available for comparison, laboratories had correctly reported parasite densities above 2% in 87.5% (70/80) samples; they had incorrectly reported parasite densities > 2% in an extra 52 (8.9%) samples. Laboratories had correctly reported P. falciparum schizonts and gametocytes in 25.6% (11/43) and 56.7% (17/30) samples, respectively. Conclusion Diagnostic laboratories in a malaria ...