The pulmonary artery catheter (PAC)-despite its invasiveness-remains the gold standard for cardiac output (CO) monitoring. The FloTrac system, a less invasive hemodynamic monitor has been developed, which estimates CO using arterial pressure waveform analysis without external calibration. Recently, an upgraded version of FloTrac system with improved algorithm to follow changes in vascular resistance was introduced into the market. The aim of this study was to assess the reliability of the CO estimated from the fourth-generation FloTrac/EV1000 system (COFT) compared to that measured with PAC using the thermodilution method (COPAC) during robotic-assisted off-pump coronary artery bypass (OPCAB) surgery. COFT and COPAC were obtained simultaneously at 4 predefined time points during robotic-assisted OPCAB: 5min after the induction of general anesthesia (T1), after starting one-lung ventilation (T2), after capnothorax (T3), and after mini-thoracotomy was performed (T4). The agreement of data was investigated by Bland-Altman analysis. Thirty-four patients were initially enrolled. After exclusion, 32 patients and a total of 128 paired CO measurements were obtained. The overall bias was 1.46L/min, the 95% limits of agreements were - 3.40 to 6.33L/min, and the percentage error was 72.98%. Regression analysis of the systemic vascular resistance index (SVRI) and the bias between COPAC and COFT showed that the bias was moderately correlated with the SVRI (r2 = 0.43; p < 0.0001). Despite a software upgrade, the reliability of the fourth-generation FloTrac/EV1000™ system during robotic-assisted OPCAB to estimate CO was not acceptable, especially in patients with low SVRI.