Dianthus broteri is an endemic complex which is considered the largest polyploid series within the Dianthus genus. This polyploid species involves four cytotypes (2×, 4×, 6× and 12×) with spatial and ecological segregation. The study of gene expression in polyploid species must be very rigorous because of the effects of duplications on gene regulation. In these cases, real-time polymerase chain reaction (qPCR) is the most appropriate technique for determining the gene expression profile because of its high sensitivity. The relative quantification strategy using qPCR requires genes with stable expression, known as reference genes, for normalization. In this work, we evaluated the stability of 13 candidate genes to be considered reference genes in leaf and petal tissues in Dianthus broteri. Several statistical analyses were used to determine the most stable candidate genes: Bayesian analysis, network analysis based on equivalence tests, geNorm and BestKeeper algorithms. In the leaf tissue, the most stable candidate genes were TIP41, TIF5A, PP2A and SAMDC. Similarly, the most adequate reference genes were H3.1, TIP41, TIF5A and ACT7 in the petal tissue. Therefore, we suggest that the best reference genes to compare different ploidy levels for both tissues in D. broteri are TIP41 and TIF5A.