Measurement of an internal control gene to adjust for sample and assay variation is a key step for successful quantitative RNA assays. Objective of this study was to select an appropriate internal control for real-time polymerase-chain reaction (PCR) assays. After titration for their PCR efficiency, all PCR primers of HPRT, 36B4, b2MG and bAct were used to evaluate expression consistency in hepatic biopsies from Holstein heifers in positive and negative energy balance. Using coefficient of variation of candidate gene overall average Ct, 36B4 and HPRT were suggested based on their relative consistency. In addition, computer program geNorm and Norm Finder were used to assist control gene selection, and both programs identified HPRT and b2MG were suitable control. With all samples and genes of interest in real-time RT-PCR, HPRT performed better than other samples as internal reference in our quantitative RNA assays. Result suggested that effects of specific experimental conditions on expression of the internal control should be evaluated for every study rather than assuming a consistent expression under all experimental conditions.