Abstract. To study gene expression and to determine distinctive characteristics of embryos produced by different methods, normalisation of the gene(s) of interest against reference gene(s) has commonly been employed. Therefore, the present study aimed to assess which reference genes tend to express more stably in single porcine blastocysts produced in vivo (IVO) or by parthenogenetic activation (PA), in vitro fertilisation (IVF) and somatic cell nuclear transfer (SCNT) using different analysis programs, namely geNorm, Normfinder and Bestkeeper. Commonly used reference genes including 18S rRNA (18S), H2A histone family, member Z (H2A), hypoxanthine phosphoribosyltransferase1 (HPRT1), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), ribosomal protein 4 (RPL4), peptidylprolyl isomerase A (PPIA), beta actin (ACTB), succinate dehydrogenase complex, subunit A (SDHA) and hydroxymethylbilane synthase (HMBS2) were analysed; most of them resulted in significantly (P , 0.05) different cycle threshold (CT) values in porcine embryos except for SDHA and H2A. In evaluation of stable reference genes across in vivo and in vitro porcine blastocysts, three kinds of programs showed slightly different results; however, there were similar patterns about the rankings of more or less stability overall. In conclusion, SDHA and H2A were determined as the most appropriate reference genes for reliable normalisation in order to find the comparative gene expression in porcine blastocysts produced by different methods, whereas 18S was regarded as a less-stable reference gene. The present study has evaluated the stability of commonly used reference genes for accurate normalisation in porcine embryos to obtain reliable results.