Mean arterial pressure (MAP) is determined in most automated oscillometric blood pressure devices, but its derivation has been little studied. In this research, different techniques were studied and compared with the auscultatory technique. Auscultatory systolic and diastolic blood pressure (SBP and DBP) were obtained in 55 healthy subjects by two trained observers, and auscultatory MAP was estimated. Automated MAP was determined by six techniques from oscillometric cuff pressures recorded digitally and simultaneously during manual measurement. MAPs were derived from the peak and foot of the largest oscillometric pulse, and from time domain curves fitted to the sequence of oscillometric pulse amplitudes (4th order and three versions of the 6th order polynomial curve). The agreement between automated and auscultatory MAPs was assessed. Compared with the auscultatory MAP, the automated MAP from the baseline cuff pressure at the peak of the 6th order polynomial curve had the smallest mean paired difference (-1.0 mmHg), and smallest standard deviation of paired differences (3.7 mmHg). These values from the peak of the largest oscillometric pulse were -1.3 and 6.2 mmHg, respectively. Determining MAP from a model of the oscillometric pulse waveform had the smallest differences from the manual auscultatory technique.