The main purpose was to investigate whether the perception of effort during the two first repetitions of strength exercises could be an adequate strategy for estimating the strength-training zone. The sample comprised 11 women (18 to 35 years-old). In the first week, the volunteers performed a 1-RM test in seven exercises on strength machines, and the load was calculated to reach 50%, 70% and 90% of the 1-RM. Over the next three weeks, the volunteers were required to perform randomly the exercises at these three intensities. After the two first repetitions, the volunteers were questioned about how many repetitions they believed they could achieve until failure (self-estimated). Additionally, volunteers were asked to indicate their exertion according Borg scale. After volunteers performed every exercise until concentric failure to complete the repetition maximum test (RMs test). The data were analyzed using linear regression, Pearson correlation and paired t-test. The results showed that the self-estimated number of repetitions underestimated 44% and 30% of the mean values of repetition maximum obtained directly at intensities of 50% and 70% (p < 0.05), respectively. Although repetition maximum were correlated with Borg scale (r = -0.23 to -0.41; p < 0.05) and self-estimated number of repetitions (r = 0.25 to 0.41; p < 0.05), the standard errors of estimate obtained by linear regression were very high (40% to 49%), which prevented any estimation equations. In conclusion, the perception of effort during the two first repetitions is not a satisfactory strategy for estimating the strengthtraining zone.