This study assessed the construct validity of accelerometry-derived net force to quantify the external demands of basketball movements. Twenty-eight basketballers completed the Yo-Yo intermittent recovery test (Yo-Yo-IR1) and basketball exercise simulation test (BEST). Intensity was quantified using accelerometry-derived average net force (AvFNet) and PlayerLoadTM per minute (PL/min). Within-player correlations were determined between intensity and running speed during Yo-Yo-IR1. Measured AvFNet was determined for movements during the BEST and predicted AvFNet was calculated using movement speed and correlations from Yo-Yo-IR1. Relationships between AvFNet and running speed during Yo-Yo-IR1 were nearly perfect (r2=0.95, 95% CI: 0.94–0.96; p<0.001) and stronger than correlations between running speed and PL/min (r2=0.80, 95% CI: 0.73–0.87; p<0.001). Differences between measured and predicted AvFNet were small during jogging and running (<1%), but large for basketball movements including jumping, change-of-direction and shuffling (15%–41%). As hypothesised, AvFNet differed by playing position (11%–16%; p<0.001) and reflected the additional demand upon players with larger body mass and lower movement efficiency. Both sprint speed and AvFNet reduced during the course of the BEST (p≤0.013). These findings confirm the construct validity of AvFNet to quantify the external demand of basketball movements. Accelerometry-derived net force has the potential to quantify the external demands of basketballers during training and competition.