Drosophila melanogaster is one of the most important genetic models and techniques such as reverse transcription quantitative real-time PCR (RT-qPCR) are being employed extensively for deciphering the genetics basis of physiological functions. In RT-qPCR, the expression levels of target genes are estimated on the basis of endogenous controls. The purpose of these reference genes is to control for variations in RNA quantity and quality. Although determination of suitable reference genes is essential to RT-qPCR studies, reports on the evaluation of reference genes in D. melanogaster studies are lacking. We analyzed the expression levels of seven candidate reference genes (Actin, EF1, Mnf, Rps20, Rpl32, Tubulin and 18S) in flies that were injured, heat-stressed, or fed different diets. Statistical analyses of variation were determined using three established software programs for reference gene selection, geNorm, NormFinder and BestKeeper. Best-ranked references genes differed across the treatments. Normalization candidacy of the selected candidate reference genes was supported by an analysis of gene expression values obtained from microarray datasets available online. The differences between the experimental treatments suggest that assessing the stability of reference gene expression patterns, determining candidates and testing their suitability is required for each experimental investigation.