This study investigated the etiology and recovery of knee extensor muscle fatigue following simulated basketball match-play. Thirteen adult male competitive basketball players (age: 25 ± 4 years, stature: 185 ± 9 cm, body mass: 86 ± 14 kg, body fat: 13 ± 4%) completed a simulated match-play (i.e., the Basketball Activity Simulation Protocol) consisting of standardized specific match-based basketball activities. Before (PRE) and immediately after (POST) the match-play, the neuromuscular function of the knee extensors was evaluated to determine the amount of muscle fatigue and its origin. Assessments were also repeated 24 h (POST24) and 48 h (POST48) after the match-play to evaluate muscle fatigue recovery. The main outcomes were maximal voluntary contraction (MVC) torque, voluntary activation estimated through superimposed stimuli, electrically-evoked twitch and doublet peak torque (PT), and the 10:100 Hz doublet ratio. The Total Quality Recovery (TQR) scale was used to assess the perceived recovery status at PRE, POST24 and POST48. Time-related changes (oneway repeated-measures ANOVA) were observed for MVC torque (main effect: P = 0.002, moderate; post hoc: POST < PRE, small), twitch and doublet PT (P < 0.001, strong; POST < PRE, POST24 and POST48, moderate-to-large) and 10:100 Hz doublet ratio (P < 0.001, strong; POST < PRE, POST24 and POST48, large; POST48 > PRE, moderate). Voluntary activation and TQR were not affected at the different time-points (P = 0.060 and P = 0.455, minimum, respectively). In conclusion, basketball match-play significantly reduced knee extensor MVC strength, with baseline levels being restored within 24 h. Muscle fatigue was accompanied by a significant pre-to-post match reduction of electrically-evoked torque responses, indicative of peripheral fatigue (and evidence of low-frequency fatigue), while no signs of central fatigue were noted.