Fluorescence in-situ hybridization (FISH) for application in preimplantation genetic diagnosis (PGD) of aneuploidy has been used successfully, but stringent scoring criteria to score FISH signals have not been developed. In the present study a FISH protocol to simultaneously enumerate chromosomes X, Y, 13, 16, 18, and 21 was used to evaluate two different scoring criteria. The criteria consider hybridization signal size, shape, and vicinity to other signals and nuclear diameter. For this purpose, 74 embryos (412 blastomeres) donated for research had most or all of their cells analysed. The least error-prone criterion (9%) was selected for use in PGD cases. Some probes produced more errors than others, and these criteria may provide clues to improve these probes. The same probe solution was applied to 55 PGD cases and a total of 307 embryos. Of the non-transferred embryos, 67 were fully reanalysed and 1.5% (1/67) of them were falsely diagnosed as normal, while 19% (13/67) were falsely diagnosed as abnormal. Twelve of the patients became pregnant after PGD.