Polycystic ovary syndrome (PCOS) is the most prevalent ovulatory and endocrine disorder affecting reproductive-aged women, yet the absence of a specific, rapid molecular diagnostic marker results in diagnostic delays and inaccuracies. Given the critical role of RNA modifications in disease pathology, this study utilized a high-throughput RNA modification profiling platform to investigate 15 types of peripheral blood RNA modification patterns in individuals with ovulatory disorders, including PCOS and primary ovarian insufficiency (POI), and control subjects. Our results revealed that distinct modification profiles correspond to specific disease states, with significant shifts in RNA modification inter-correlations observed across conditions. Additionally, specific RNA modifications were associated with clinical features, such as serum levels of testosterone and the follicle number per ovary (FNPO). To optimize diagnostic precision, we evaluated various machine learning models, identifying that combining mA and mG modifications in a light gradient boosting machine model (LightGBM) achieves the highest accuracy in distinguishing PCOS, outperforming traditional diagnostic markers. This highlights the potential of RNA modification profiling as a novel, high-accuracy diagnostic tool for PCOS in clinical settings.