Quantitative real-time PCR (qPCR) is currently the most accurate method for detecting differential gene expression, but depends greatly on normalization with stably expressed housekeeping genes. Transcriptomics analyses and experimental validation in different plant species have shown that the reliability of these endogenous controls can be influenced by, growth conditions and organs/tissues examined. Thus, reliable validation of reference genes is required to ensure proper normalization. This paper presents a systematic comparison of ten potential reference genes in sunflower: five commonly used genes (Actin, Elongation Factor1, Plastid-encode RNA polymerase, Tubulin, and Ubiquitin, as ACT, EF1, PEP, TUB, and UBQ respectively), as well as five new candidates (Translation initiation factor, MicroRNA precursors 171 and 156, Ask-interacting protein, and Protein of unknown function, as ETIF5, MIR171, MIR156, SKIP, and UNK2 respectively). Reference gene expression stability was examined by qPCR across 20 biological samples, representing different tissues at various developmental stages. Expression of all 10 genes was variable to some extent, but that of ACT, UNK2, and EF1 was overall the most stable. A combination of ETIF5/UNK2/EF1 would be appropriate to use as a reference panel for normalizing gene expression data among vegetative tissues, whereas the combination of ACT/MIR156/UNK2 is most suitable for reproductive tissues. Reference genes selected in this study were further validated by examining relative expression of ahas1, one of three acetohydroxyacid synthase genes of sunflower. Our identification and validation of suitable normalizer genes will be of use to ensure accurate results in future transcriptomics studies in this crop.