This study aimed to compare force, velocity, and power output collected under different loads, as well as the force-velocity (F-V) relationship between three measurement methods. Thirteen male judokas were tested under four loading conditions (20, 40, 60, and 80 kg) in the countermovement jump (CMJ) exercise, while mechanical output data were collected by three measurement methods: the Samozino's method (SAM), a force platform (FP), and a linear velocity transducer (LVT). The variables of the linear F-V relationship (maximum force [F0], maximum velocity [V0], F-V slope, and maximum power [P0]) were determined. The results revealed that (1) the LVT overestimated the mechanical output as compared to the SAM and FP methods, especially under light loading conditions, (2) the SAM provided the lowest magnitude for all mechanical output, (3) the F-V relationships were highly linear either for the SAM (r = 0.99), FP (r = 0.97), and LVT (r = 0.96) methods, (4) the F-V slope obtained by the LVT differed with respect to the other methods due to a larger V0 (5.28 +/- 1.48 m.s(-1)) compared to the SAM (2.98 +/- 0.64 m.s(-1)) and FP (3.06 +/- 0.42 m.s(-1)), and (5) the methods were significantly correlated for F0 and P0, but not for V0 or F-V slope. These results only support the accuracy of the SAM and FP to determine the F-V relationship during the CMJ exercise. The very large correlations of the SAM and LVT methods with respect to the FP (presumed gold-standard) for the mean values of force, velocity and power support their concurrent validity for the assessment of mechanical output under individual loads.