Persistently Infected (PI) animals play a central role in the transmission of BVDV infection between cattle herds. Thus, promoting the certification of non-PI animals is a relevant approach for improving control, as it contributes to securing the trade. The objectives of this study were: i) to assess the reliability of diverse certification criteria, and ii) to identify risk factors for erroneous certification. To do so, the proportion of animals wrongly certified as non-PI on the basis of tests performed after the certification date, was calculated for each criterion. The data used were collected in herds located in Brittany, involved in either a clearance process for those that were infected, or in a surveillance process for herds that were BVDV-free. A total of 23 criteria were defined by combining the technical characteristics of the tests (individual vs. pool; single vs. repeated; direct vs. indirect tests), and some pathogenic characteristics of BVDV infection. Overall, the rates of wrongly-certified animals were low (mean: 1.3 10−4). Direct and indirect criteria had equivalent performances. Heifers from birth, and even foetuses in the last third of gestation, are certified, provided that the herd to which they belong has been free of BVDV for more than 2.5 years. The risk for wrong certification increased in the case of PIs present in the herd or its surroundings. The simplicity of the output-based approach described here, and the excellent performance of indirect criteria relying on serological monitoring of BTM, make it particularly interesting, as its use could facilitate trade between countries.