The aim of the present study was to present the reliability and normative values of wearable inertial sensors kinematics during football-specific cut maneuver tasks collected on the football field for assessing anterior cruciate ligament (ACL) injury risk. Forty-seven academy football players (age 15.9 ± 2.4 years, female n = 20) performed a planned 90° change of direction within the Agility T-test and unplanned football-specific changes of direction (FS deceiving action). Kinematics was collected through eight wearable inertial sensors (100 hz, MTw Awinda, Movella). Intraclass correlation coefficient and Root Mean Square Error were used to inspect test-retest and side-to-side reliability of peak and waveform kinematics. Normative kinematics was compared between male and female players (t-test with Cohen's d, p < 0.05). Test-retest reliability was moderate-to-excellent in most of the parameters (r=0.40-0.92). Side-to-side reliability was worse than test-retest (both movement tasks). Female players showed worse movement quality than males with greater peak values on the frontal and transverse planes at the knee, pelvis and trunk and smaller knee and trunk flexion (d = 0.50-1.1 in Agility T-test, = 0.39-0.73 in FS deceiving action). The on-field cut maneuver kinematics by wearable sensors demonstrated sufficient reliability for most joints. Reliability and normative values might help to objectify ACL injury prevention programs in football academies.