AIMS: Bicuspid aortic valve (BAV) is the most common congenital heart defect, affecting at least 2% of the population. The embryonic origins of BAV remain poorly understood, with few assays for validating patient variants, limiting the identification of causative genes for BAV. In both human and mouse, the left and right leaflets of the arterial valves arise from the outflow tract cushions, with interstitial cells originating from neural crest cells and the overlying endocardium through endothelial-to-mesenchymal transition (EndoMT). In contrast, an EndoMT-independent mechanism of direct differentiation of cardiac progenitors from the second heart field (SHF) is responsible for the formation of the anterior and posterior leaflets. Defects in either of these developmental mechanisms can result in BAV. Although zebrafish have been suggested as a model for human variant testing, their naturally bicuspid arterial valve has not been considered suitable for understanding human arterial valve development. Here, we have set out to investigate to what extent the processes involved in arterial valve development are conserved in zebrafish and, ultimately, whether functional testing of BAV variants could be carried out. METHODS AND RESULTS: Using a combination of live imaging, immunohistochemistry, and Cre-mediated lineage tracing, we show that the zebrafish arterial valve primordia develop directly from SHF progenitors with no contribution from EndoMT or neural crest, in keeping with the human and mouse anterior and posterior leaflets. Moreover, once formed, these primordia share common subsequent developmental events with all three aortic valve leaflets. CONCLUSION: Our work highlights a conserved ancestral mechanism of arterial valve leaflet formation from the SHF and identifies that development of the arterial valve is distinct from that of the atrioventricular valve in zebrafish. Crucially, this confirms the utility of zebrafish for understanding the development of specific BAV subtypes and arterial valve dysplasia, offering potential for high-throughput variant testing.