Background Coronary microvascular function can be assessed by transthoracic Doppler echocardiography as a coronary flow velocity reserve (TTDE CFVR) and by positron emission tomography as a myocardial blood flow reserve (PET MBFR). PET MBFR is regarded the noninvasive reference standard for measuring coronary microvascular function but has limited availability. We compared TTDE CFVR with PET MBFR in women with angina pectoris and no obstructive coronary artery disease and assessed repeatability of TTDE CFVR. Methods From a cohort of women with angina and no obstructive coronary artery stenosis at invasive coronary angiography, TTDE CFVR by dipyridamole induced stress and MBFR by rubidium-82 PET with adenosine was successfully measured in 107 subjects. Repeatability of TTDE CFVR was assessed in 10 symptomatic women and in 10 healthy individuals. Results MBFR was systematically higher than CFVR. Median MBFR (interquartile range, IQR) was 2.68 (2.29–3.10) and CFVR (IQR) was 2.31 (1.89–2.72). Pearson's correlation coefficient was 0.36 (p < 0.01). Limits of agreement (2·standard deviation) assessed by the Bland–Altman (confidence interval, CI) method was 1.49 (1.29;1.69) and unaffected by time-interval between examinations. Results were similar when adjusting for rate pressure product or focusing on perfusion of the left anterior descending artery region. Limits of agreement (CI) for repeated CFVR in 10 healthy individuals and in 10 women with angina was 0.44 (0.21;0.68) and 0.48 (0.22; 0.74), respectively. Conclusion CFVR had a good repeatability, but the agreement between CFVR and MBFR was modest. Divergence could be due to methodology differences; TTDE estimates flow velocities whereas PET estimates myocardial blood flow.