Fluorescence in situ hybridization (FISH) is a cytogenetic technology used to detect chromosomal abnormalities in preimplantation human embryos. However, its efficiency is not stable due to improper sample preparation. The present study was designed to modify the current sample preparation technique and then to evaluate its efficiency in human preimplantation genetic diagnosis (PGD). Day 3 cleavage embryos as well as day 5 and 6 blastocysts were biopsied by mechanical aspiration method. In the present study, two methods were used for sample preparation of the biopsied cells. Method I was the traditional method, in which each blastomere was placed in a hypotonic solution for 5 min and then fixed on glass slides. The slides were kept at room temperature before the FISH procedures. Method II was a modified method, in which all blastomeres were placed individually in hypotonic solution drops covered by oil for at least 5 min and then fixed on slides with 0.1% Tween/HCl. After fixation, the slides were kept at –20°C for at least 30 min before the FISH procedures. The two methods were compared in terms of time consumption and proportions of blastomeres with FISH signals. In total, 329 blastomeres from day 3 embryos were fixed by Method I with an average fixation time of 8–10 min for each blastomere. By contrast, with Method II, 362 blastomeres were fixed and the average time was 3–4 min for each blastomere. After FISH, more nuclei had signals with Method II (97.2%) than with Method I (86.9%). All cells that were biopsied from blastocysts and prepared with Method II had FISH signals. However, Method I was not suitable for the fixation of multiple cells biopsied from blastocysts as cells were not traceable during the fixation. The present study indicates that proper sample preparation is critical for obtaining FISH signals in cells biopsied from preimplantation human embryos; hence these modifications can increase the efficiency of human PGD.