The assessment of teratogenic effects of chemicals is generally performed using in vivo teratogenicity assays e.g., in rats or rabbits. Following the 3R principles, the development of alternative methods is encouraged to reduce the number of animal tests. From this perspective, we have developed an in vitro assay (mDarT) using the zebrafish Danio rerio embryo teratogenicity assay (DarT) combined with an exogenous mammalian metabolic activation system (MAS), able to biotransform proteratogenic compounds. Cyclophosphamide, ethanol, benzo[a]pyrene and thalidomide were used as test materials to assess the efficiency of this assay. Briefly, the zebrafish embryos were co-cultured at 2 hpf (hours post fertilization) with the test material at varying concentrations, mammalian liver microsomes from different species and NADPH for 60 min at 32°C under moderate agitation in Tris buffer. The negative control (test material alone) and the MAS control (MAS alone) were incubated in parallel. For each test group, 20 eggs were used for statistical robustness. Afterwards fish embryos were transferred individually into 24-well plates filled with fish medium for 48 hours at 26°C with a 12 hour-light cycle. Teratogenicity was scored after 24 and 48 hpf using morphological endpoints. The test was considered to be valid if a minimum of 90% of fish eggs developed normally for the two controls (test material alone and MAS alone). For each test material, the experiment was repeated three times with the controls satisfying the validation criteria (≤ 10% impaired embryos). Indeed, no significant teratogenic effects were observed compared to controls in fish embryos exposed to the proteratogens alone (i.e., without metabolic activation) or the MAS alone. In contrast, the four test materials induced significant abnormalities in fish embryos when co-incubated with animal liver microsomes. For cyclophosphamide, ethanol and thalidomide a concentration-response relationship was shown and the qualitative nature of the malformations was similar between fish embryos and humans. Benzo[a]pyrene was demonstrated to be significantly teratogenic in fish embryos in spite of no concentration-response and unspecific teratogenic fingerprints. We conclude that the application of animal liver microsomes will improve and refine the DarT as a predictive and valuable alternative method to screen teratogenic substances.