Human amniotic membrane and amniotic membrane-derived mesenchymal stromal cells (hAMSCs) have produced promising results in regenerative medicine, especially for the treatment of inflammatory-based diseases and for different injuries including those in the orthopedic field such as tendon disorders. hAMSCs have been proposed to exert their anti-inflammatory and healing potential via secreted factors, both free and conveyed within extracellular vesicles (EVs). In particular, EV miRNAs are considered privileged players due to their impact on target cells and tissues, and their future use as therapeutic molecules is being intensely investigated. In this view, EV-miRNA quantification in either research or future clinical products has emerged as a crucial paradigm, although, to date, largely unsolved due to lack of reliable reference genes (RGs). In this study, a panel of thirteen putative miRNA RGs (let-7a-5p, miR-16-5p, miR-22-5p, miR-23a-3p, miR-26a-5p, miR-29a-5p, miR-101-3p, miR-103a-3p, miR-221-3p, miR-423-5p, miR-425-5p, miR-660-5p and U6 snRNA) that were identified in different EV types was assessed in hAMSC-EVs. A validated experimental pipeline was followed, sifting the output of four largely accepted algorithms for RG prediction (geNorm, NormFinder, BestKeeper and ΔCt method). Out of nine RGs constitutively expressed across all EV isolates, miR-101-3p and miR-22-5p resulted in the most stable RGs, whereas miR-423-5p and U6 snRNA performed poorly. miR-22-5p was also previously reported to be a reliable RG in adipose-derived MSC-EVs, suggesting its suitability across samples isolated from different MSC types. Further, to shed light on the impact of incorrect RG choice, the level of five tendon-related miRNAs (miR-29a-3p, miR-135a-5p, miR-146a-5p, miR-337-3p, let-7d-5p) was compared among hAMSC-EVs isolates. The use of miR-423-5p and U6 snRNA did not allow a correct quantification of miRNA incorporation in EVs, leading to less accurate fingerprinting and, if used for potency prediction, misleading indication of the most appropriate clinical batch. These results emphasize the crucial importance of RG choice for EV-miRNAs in hAMSCs studies and contribute to the identification of reliable RGs such as miR-101-3p and miR-22-5p to be validated in other MSC-EVs related fields.