Objective: To verify the influence of 3 potential factors on the accuracy of non-invasive central BP measurement, namely the calibration peripheral BP (brachial or radial), calibration methods (systolic/diastolic BP or mean/diastolic BP), and devices. Design and method: 50 participants (aged 62.4 ± 8.9 years) without overt heart diseases were recruited from December 2017 to July 2018. Invasive aortic BP were measured with 5F Pigtail (Cordis, U.S.) and invasive radial BP were measured with 6F Radial sheath (Cordis, U.S.) after connection to the recorder (St. Jude Medical, U.S.). Brachial BP were measured non-invasively (HEM-8102A, Omron, Japan). At the same time, non-invasive central BP were measured with 3 devices, namely SphygmorCor (AtCormedical, Australia), Mobil-O-Graph (IEM, Germany), and PulsePen (DiaTecne, Italy). Non-invasive central BP were calibrated with different peripheral BP or methods manually when available.Results: Of all participants, 23 were men and 31 were hypertensive patients. The mean invasive aortic BP was 127 ± 19/95 ± 14 mmHg. Correlation analysis showed that invasive central SBP (cSBP) significantly correlated with all non-invasive cSBP (P < 0.001), no matter which device or calibration strategy was. In ANOVA analyses, recordings from Mobil-O-Graph device were significantly lower than invasive measurements (P = 0.04), while other recordings were similar to invasive measurements (P>=0.89). Next, all recordings were aggregated and re-classified based on their calibration peripheral BP, calibration methods, and devices, respectively. Taking invasive measurements as the reference, it was shown that calibration with non-invasive brachial BP were significantly lower than the reference (P = 0.01), and calibration with invasive radial BP was not different from the reference (P = 0.10). Calibration method, either SBP/DBP or mean BP/DBP, did not significantly influence the non-invasive cSBP values (P> = 0.24). Though Mobil-O-Graph had significantly lower recordings of cSBP than the reference, it was not significantly different from the SphygmorCor Device (P = 0.13). Multiple regression analysis (including age and gender as confounders) showed that, gender and calibration peripheral BP could significantly influence the accuracy of non-invasive cSBP measurement. Conclusions: The use of radial BP rather than brachial BP might be better in non-invasive cSBP measurement, and the Mobil-O-Graph device is not recommended. Calibration methods have little influence on cSBP measurement.