A repeatable and reliable reverse transcription quantitative PCR (RT-qPCR) experiment depends upon proper reference genes (RGs) selection. This study aims to examine the expression stability of nine candidate RGs for the Avena sativa - Blumeria graminis experimental setup. B. graminis causes powdery mildew - the most devastating and economically important fungal disease of crops worldwide. RGs were evaluated in Pm3 and Pm4 oat differential lines and the susceptible cultivar Fuchs during compatible and incompatible interactions with different pathotypes of Blumeria graminis f. sp. avenae in six-time points post inoculation. The identification of genes exhibiting high expression stability was done by four algorithms (geNorm, NormFinder, BestKeeper and deltaCt). The results indicated that regardless of the analysed group, two most stable RGs are required for data normalization. The most sufficient RGs combination was HNR (heterogeneous nuclear ribonucleoprotein 27 C) + EIF4A (eukaryotic initiation factor 4 A‑3). ARF (ADP‑ribosylation factor) could also be pondered as demonstrating high expression stability. These genes can be considered universal candidates for RT-qPCR normalization to study interaction with B. graminis as well as Puccinia coronata and Puccinia graminis, as confirmed by our previous research. The worst candidate for data standardisation was TUA (α- tubulin). To our best knowledge, this is the first report regarding RGs' selection in this pathosystem. Identified RGs are proper normalisation candidates for gene expression studies in the A. sativa infected by B. graminis as well as other related pathogens.