Reverse transcription quantitative real-time polymerase chain reaction (qRT-PCR) is a commonly used technology for gene expression and transcriptome analysis. Normalization is a process that is necessary to accurately analyze qRT-PCR data. Stability of reference gene expression is required for this process. Due to the large variation in expression levels of reference genes obtained from different experimental conditions, gene expression stabilities must be evaluated and identified in all experimental systems. In the present paper, the stability of the expression levels of seven potential reference genes in pepper are assessed using qRT-PCR analysis to determine optimal reference genes. These reference genes are evaluated in different pepper tissues, abiotic stress, and hormonal treatment samples. Three common statistical algorithms, geNorm, NormFinder, and BestKeeper, are used to identify expression stability and provide an accurate selection of reference genes. Two reference genes, beta tubulin and ubiquitin-conjugating protein (UBI-3), showed high stability in sample pools with abiotic stress and hormonal treatments. Among the sample pools tested, UBI-3 and glyceraldehyde-3-phosphate dehydrogenase expression levels were the most stable in different tissues. Therefore, these reference genes are selected for qRT-PCR analysis under the experimental conditions tested in pepper. In contrast, ubiquitin-conjugating enzyme and actin genes are identified as the least stable reference genes in all the groups tested, confirming that they are not suitable for normalization. Validation of these candidate genes could provide useful guidelines for reference gene selection in qRT-PCR studies in pepper. Crown Copyright (C) 2011 Published by Elsevier Inc. All rights reserved.