Quantitative reverse transcription PCR (RT-qPCR) is a powerful molecular technique that enables gene expression studies to be performed on extremely small samples such as oocytes and pre-implantation embryos. However, the high sensitivity of this technique requires that to prevent bias in expression data interpretation, the reference genes used for normalization are fully validated before use. Difficulties in choosing a reference gene are further compounded by the variable RNA content of the maturing oocyte. In the present study, we evaluated eight commonly used reference genes such as ACTB, GAPDH, H2AFZ, HPRT1, PPIA, SDHA, TUBB and YWHAZ and for sheep oocyte RT-qPCR before and after in vitro maturation. We have also compared different cDNA priming strategies using random hexamers or oligo-dT. GeNorm analysis of the results identified the most reliable genes for normalization to be SDHA, TUBB and PPIA when oocyte cDNA was made with random hexamers, and YWHAZ, TUBB and SDHA when oligo-dT primers were used (H2AFZ and HPRT1 were excluded from the geNorm analysis). Interestingly, the analysis revealed that the least stable genes were ACTB and GAPDH, which are the conventional 'housekeeping' genes used in many studies. We recommend the use of three reference genes to calculate a normalization factor to accurately quantify transcript abundance in sheep oocytes and these vary with the cDNA priming strategy employed.