This study examined the potential use of myoelectric signal analysis for determination of anaerobic threshold (AT). Ten college students performed incremental exercise on a bicycle ergometer during which blood samples from either the brachial or radial artery were drawn and corresponding intergrated electromyogram (IEMG) and power spectra also were recorded. Blood AT was determined by observing the changes in arterial blood lactate, Po2 Pco2, HCO3-, and pH. EMG AT was determined by the non-linear increase in IEMG and abrupt increase in frequency band width at 70% of the peak frequency (FRQ70%) which could represent the most active frequency band. The results indicated that a significant increase in FRQ70% occurred immediately after the occurrence of blood AT (p less than 0.05). However, no significant increase could be found in peak frequency. Regression analyses revealed that EMG AT Vo2 correlated significantly with blood AT Vo2 (r = 0.921, p less than 0.001) and with gas exchange AT Vo2 (r = 0.921, p less than 0.001). It was concluded that analysis of myoelectric signals, which reflect motor unit recruitment and discharge frequency, could provide a non-invasive estimate of AT.