AbstractThe aim of this study was to assess whether lifting velocity (MV) can provide accurate estimations of the maximum number of repetitions that can be performed to failure (RTF) during the parallel back‐squat exercise performed in a Smith machine. Thirty male wrestlers from the Turkey Olympic preparation center (age = 22.6 ± 2.2 years) completed four testing sessions: a session to determine the back‐squat one‐repetition maximum [1RM], two sessions consisting of single sets to failure against three loads (90%‐80%‐70%1RM), and one session consisting of four sets to failure against the 75%1RM. The goodness‐of‐fit of the generalized RTF‐MV relationship was strong (r2 = 0.838), but the individualized RTF‐MV relationships were stronger (r2 = 0.957 ± 0.058). Only 3 out of 60 individualized RTF‐MV relationships revealed a r2 lower than the r2 of the generalized RTF‐MV relationship (r2 = 0.685, 0.779 and 0.810). The reliability of the fastest MV associated with each RTF ranged from acceptable (4 out of 15 RTFs) to high (11 out of 15 RTFs). The raw and absolute errors in the prediction of RTF did not increase under fatigue and were comparable for both generalized (raw errors: −1.0–0.3 repetitions; absolute errors: 1.1–1.7 repetitions) and individualized (raw errors: −0.8 to 0.1 repetitions; absolute errors: 1.2–1.8 repetitions) RTF‐MV relationships. These results indicate that RTF can be predicted with acceptable precision from MV recordings in resistance‐trained skilled wrestlers during the parallel back‐squat exercise performed in a Smith machine.