The objective of the study was to investigate the validity of the rating of perceived exertion threshold (T RPE ) for anaerobic threshold (AT) estimation during an incremental test (IT) on a cycle ergometer. Nineteen physically active and healthy men performed an IT on a cycle ergometer. The intensity corresponding to the AT was considered as the point immediately before the abrupt increase in blood lactate concentration ([La]) in the IT. T RPE was determined in the last intensity corresponding to perceptual response previous the 15 arbitrary units (AU) in the 15-point RPE scale. The heart rate (HR), [La], and RPE were obtained at the end of each stage of the IT. There was no significant difference between workload obtained from the watts) and T RPE (150 [120-150] watts) (p . 0.05). In addition, there were no differences between HR (AT, 139.5 6 12.7 b•min 21 vs. T RPE , 141.9 6 14.6 b•min 21 , p . 0.05) and [La] (AT, 3.1 [2.8-3.2] mM vs. T RPE , 3.1 [2.9-3.7] mM, p . 0.05). There was a strong correlation between the intensities in AT and T RPE during the IT (r 5 0.88). A substantial reliability (intraclass correlation coefficient 5 0.84 [0.64-0.94 confidence interval 95%]) was evidenced, as well as a substantial agreement between the AT and T RPE intensities (Pc 5 0.84), as confirmed by the Bland-Altman plot (24.7 [2 34.2/24.7]). T RPE was a predictor of the AT (b 5 1.146, R 5 0.770, SEE 5 15.070, p , 0.01). It can be concluded that T RPE is valid for AT estimation during an IT on a cycle ergometer.