The present study investigated whether or not verbal instruction affects the electromyographic (EMG) amplitude of back-squat prime movers. Fifteen resistance-trained men performed back-squat at 50%1-RM and 80%1-RM and received external (EF) or internal focus (IF) on lower-limb posterior muscles. EMG amplitude of gluteus maximus, biceps femoris, gastrocnemius medialis, vastus lateralis and tibialis anterior was recorded during both concentric and eccentric phase. During the concentric phase, the gluteus maximus and biceps femoris EMG amplitude was greater in IF vs EF at 50% [effect size (ES): 0.63(95%CI 0.09/1.17) and 0.49(0.10/0.78) respectively] and 80% [ES: 1.30(0.29/2.21) and 0.59(0.08/1.10)]. The gastrocnemius medialis EMG amplitude was greater in IF vs EF during the eccentric phase at 50% [ES: 0.73(0.13/1.33)] and at 80% [ES: 0.72(0.10/1.34)]. Concomitantly, vastus lateralis EMG amplitude was lower at 50% [ES: -0.71(-1.38/-0.04)] and 80% [ES: -0.68(-1.33/-0.03)]. During the eccentric phase, the tibialis anterior EMG amplitude was greater in IF vs EF at 50% [ES: 0.90(0.12 to 1.68)] and 80% [ES: 0.74(0.13/1.45)]. Irrespective of the load, in the thigh muscles the internal focus promoted a different motor pattern, increasing the hip extensors and reducing the knee extensor excitation during the concentric phase. Concomitantly, both ankle muscles were more excited during the eccentric phase, possibly to increase the anterior-posterior balance control. The internal focus in back-squat seems to have phase-dependent effects, and it is visible at both moderate and high load.