The objective of the study was to assess the use of maximum (V_max) and final propulsive phase (FPV) bar velocity to predict jump height in the weighted jump squat. FPV was defined as the velocity reached just before bar acceleration was lower than gravity (-9.81 m·s^-2). Vertical jump height was calculated from the take-off velocity (V_take-off) provided by a force platform. Thirty swimmers belonging to the National Slovenian swimming team performed a jump squat incremental loading test, lifting 25%, 50%, 75% and 100% of body weight in a Smith machine. Jump performance was simultaneously monitored using an AMTI portable force platform and a linear velocity transducer attached to the barbell. Simple linear regression was used to estimate jump height from the V_max and FPV recorded by the linear velocity transducer. V_max (y = 16.577x - 16.384) was able to explain 93% of jump height variance with a standard error of the estimate of 1.47 cm. FPV (y = 12.828x - 6.504) was able to explain 91% of jump height variance with a standard error of the estimate of 1.66 cm. Despite that both variables resulted to be good predictors, heteroscedasticity in the differences between FPV and V_take-off was observed (r² = 0.307), while the differences between V_max and V_take-off were homogenously distributed (r² = 0.071). These results suggest that V_max is a valid tool for estimating vertical jump height in a loaded jump squat test performed in a Smith machine.

• Vertical jump height in the loaded jump squat can be estimated with acceptable precision from the maximum bar velocity recorded by a linear velocity transducer.

• The relationship between the point at which bar acceleration is less than -9.81 m·s² and the real take-off is affected by the velocity of movement.

• Mean propulsive velocity recorded by a linear velocity transducer does not appear to be optimal to monitor ballistic exercise performance.