Aim. To study the potential of non-invasive biomarkers in the diagnosis of coronary microvascular dysfunction (CMD) and prediction of the course of heart failure with preserved ejection fraction (HFpEF) in non-obstructive coronary artery disease.Materials and methods. The 12-month observational study included 118 consecutive patients (6 patients dropped out of the study due to contact loss) with non-obstructive coronary artery disease (CAD) and HFpEF (62 [59; 64]%). At the beginning of the study, serum levels of several biomarkers were assessed using the enzyme immunoassay: N-terminal pro-B-type natriuretic peptide (NT-proBNP), vascular endothelial growth factor (VEGF), and endothelin-1. Coronary flow reserve (CFR) was examined using dynamic single photon emission computed tomography. In the absence of obstructive CAD, CMD was defined as a global decrease in CFR ≤ 2. Echocardiography was used to determine parameters of hemodynamics, LV diastolic dysfunction, and myocardial stress. LV global longitudinal strain (GLS) was assessed using 2D speckle tracking.Results. The patients were divided into groups depending on the presence of CMD: group 1 included patients with CMD (n = 43), group 2 included those without it (n = 75). In patients in group 1, serum levels of endothelin-1 were 1.9 times higher (p = 0.012), levels of VEGF were 2.16 times higher (p = 0.008), and the concentration of NT-proBNP was 2.6 times higher (p = 0.004) than in patients in group 2. According to the ROC analysis, the concentrations of endothelin-1 ≥ 6.9 pg / ml (AUC = 0.711; p = 0.040) and VEGF ≥ 346.7 pg / ml (AUC = 0.756; p = 0.002) were considered as markers associated with the presence of CMD in patients with non-obstructive CAD. The multivariate regression analysis showed that only the presence of CMD (odds ratio (OR) 2.42; 95% confidence interval (95% CI) 1.26–5.85; p < 0.001) and an increase in NT-proBNP ≥ 760.5 pg / ml (OR 1.33; 95% CI 1.08–3.19; p = 0.023) were factors associated with adverse events, and their combination increased the risk of HFpEF progression by more than 3 times (OR 3.18; 95% CI 2.76–7.98; p < 0.001), whereas markers of endothelial dysfunction were not independent predictors. Conclusion. Endothelin-1 ≥ 6.9 pg / ml and VEGF ≥ 346.7 pg / ml can be used as non-invasive markers for the diagnosis of CMD. However, markers of endothelial dysfunction were not independent predictors of HFpEF progression in patients with non-obstructive CAD during 12-month follow-up.