The pink stemborer, Sesamia inferens (Walker) (Lepidoptera: Noctuidae), has gradually become amajor threat to rice production in Asian rice growth area. Using reverse-transcription quantitative real-time PCR (qRT-PCR) to investigate gene expression of S. inferens under different conditions is a popular technique. However, numerous studies demonstrate that the experimental conditions affect the stability of reference genes. Thus, it is necessary to normalize qRT-PCR data under a specific condition using a stable reference gene as an internal control. In the present study, five algorithms (Delta Ctmethod, Bestkeeper, geNorm,Normfinder, and RefFinder) and seven candidate reference genes were evaluated to find out the most suitable reference genes under different biotic and abiotic conditions including different tissues and developmental stages, exposure to insecticides, different rice varieties, and different diets. Our results showed that the more stable series of reference genes were as follows: GAPDH, RPL10, and RPS3 for different tissues;GAPDH and RPS3 for developmental stages; RPS3 and EF1 for larval insecticide exposure; RPS3 and RPL10 for larvae reared on different rice varieties; RPS3 and ACTB for larvae fed on different diets. Hence, a combination of the best sets of reference genes for normalization of expression analysis will increase the accuracy and reliability of qRT-PCR data in S. inferens. In addition, these results establish the basis for selection of reference genes in other lepidopterous insect species.