Studying complex plant resistance mechanisms is based on transcriptome analysis using high-throughput NGS. Accurate quantification and validation of gene expression derived from in silico NGS data is performed by RT-qPCR. The aim of this research was to identify reference genes with stable expression level in order to ensure obtaining repeatable and reliable RT-qPCR data. Eleven frequently used candidate reference genes [1, 2] were evaluated in A. sativa during compatible and incompatible interactions with different pathotypes of Puccinia coronata f. sp. avenae in six time points post inoculation. The identification of genes with high expression stability was performed by four algorithms (geNorm, NormFinder, BestKeeper and ΔCt method). The results obtained confirmed that the combination of two genes would be sufficient for reliable normalization of the expression data. In general, the most stable in tested plant-pathogen system were elongation factor 1-α and heterogeneous nuclear ribonucleoprotein 27C. Eukaryotic initiation factor 4A-3 and ADP-ribosylation factor could also be considered as exhibiting high expression stability. Cyclophilin was shown by all assessment methods to be the worst candidate for normalization in this dataset. To date, this is the first report of RGs selection in A. sativa – P. coronata interaction system. The obtained results will provide valuable data necessary for a comprehensive analysis of oat gene expression in response to crown rust infection