The present investigation described and compared the internal and external training loads during basketball training. Eight semi-professional male basketball players (mean ± SD, age: 26.3 ± 6.7 years; stature: 188.1 ± 6.2 cm; body mass: 92.0 ± 13.8 kg) were monitored across a seven-week period during the preparatory phase of the annual training plan. A total of 44 total sessions were monitored. Player session ratings of perceived exertion (sRPE), heart rate (HR), and accelerometer data were collected across each training session. Internal training load was determined using the session rating of perceived exertion (sRPE), training impulse (TRIMP), and summated-heart-rate-zones (SHRZ) training load models. External training load was calculated using an established accelerometer algorithm. Pearson product-moment correlations with 95% confidence intervals (95% CI) were used to determine the relationships between internal and external training load models. Significant moderate relationships were observed between external training load and the sRPE (r42 = 0.49, 95% CI = 0.23-0.69, P <0.001) and TRIMP models (r42 = 0.38, 95% CI = 0.09-0.61, P = 0.011). A significant large correlation was evident between external training load and the SHRZ model (r42 = 0.61, 95% CI = 0.38-0.77, P < 0.001). Although significant relationships were found between internal and external training load models, the magnitude of the correlations and low commonality suggest internal training load models measure different constructs of the training process than the accelerometer training load model in basketball settings. Basketball coaching and conditioning professionals should not assume a linear dose-response between accelerometer and internal training load models during training, and are recommended to combine internal and external approaches when monitoring training load in players.