Helopeltis theivora Waterhouse is a predominant sucking pest in many tropic economic crops, such as tea, cocoa and coffee. Quantitative real-time PCR (qRT-PCR) is one of the most powerful tools to analyze the gene expression level and investigate the mechanism of insect physiology at transcriptional level. Gene expression studies utilizing qRT-PCR have been applied to numerous insects so far. However, no universal reference genes could be used for H. theivora. To obtain accurate and reliable normalized data in H. theivora, twelve candidate reference genes were examined under different tissues, developmental stages and sexes by using geNorm, NormFinder, BestKeeper, Delta Ct and RefFinder algorithms, respectively. The results revealed that the ideal reference genes differed across the treatments, and the consensus rankings generated from stability values provided by these programs suggested a combination of two genes for normalization. To be specific, RPS3A and Actin were the best suitable reference genes for tissues, RPL13A and GAPDH were suitable for developmental stages, EF1α and RPL13A were suitable for sexes, and RPL13A and RPS3A were suitable for all samples. This study represents the first systematic analysis of reference genes for qRT-PCR experiments in H. theivora, and the results can provide a credible normalization for qRT-PCR data, facilitating transcript profiling studies of functional genes in this insect.