Background: Before studying the gene expression of different organisms, it is important to determine the best reference gene. At present, the most accurate method for detecting gene expression is quantitative real-time PCR(RT-qPCR). By using this method, reference genes that are stable in different biological systems and under different conditions can be obtained. Toona ciliata Roem (T. ciliata). is a valuable and fast-growing timber species. In this study, 20 reference genes were identified through RT-qPCR, as a primary prerequisite for future gene expression analysis. Four different methods, geNorm, NormFinder, BestKeeper, and RankAggreg were used to evaluate the stability of expression of 20 candidate reference genes in various tissues under different conditions. Results: The experimental results showed that TUB-α was the most stably expressed reference gene across all samples; UBC17 was found to be the most stable in leaves & young stems under Hypsipyla robusta (H. robusta) and methyl jasmonate (MeJA) treatment. In addition, under H. robusta treatment, PP2C59 and UBC5B were the best-performing genes in leaves, while HIS1 and ACT7 were the best reference genes in young stems. Under low temperature (4℃) treatment, the two best reference genes were 60S-18 and TUB-α. The expression of HIS6 and MUB1 was the most stable under PEG6000 treatment. The accuracy of the selected reference genes was verified using transcription factor MYB3(TcMYB3) genes. Conclusions: This is the first report to verify the best reference genes for normalizing gene expression in T. ciliata under different conditions, which will facilitate the future elucidation of gene regulations in this species.