Wooden breast (WB) myopathy was investigated in broilers fed varying energy and protein at early ages. Correlation analyses were conducted between echogenicity of ultrasound images (US) of breast muscle from live birds and WB after slaughter. A total of 1,000 Cobb 500 one-day-old male chicks were fed on five dietary programs with eight replicates of 25 birds each, in a completely randomized design. Control feeds (commercially used ME and ideally balanced amino acids) or low-density feeds (low EP, with reductions of 50 kcal/kg ME and 0.20% dig. Lys compared to the control) were formulated. Feeds were provided in different periods: 1 to 7 d, 8 to 14 d, 15 to 21 d or 22 to 28 d. All broilers were fed a common basal diet thereafter until 49 d. Images using US were obtained once a week from all individuals and WB scored from one slaughtered bird per replication (0, normal; 1, mild hardening in the upper breast muscle; 2; moderate hardening in the upper and/or lower breast muscle; 3, severe hardening; 4, severe hardening with hemorrhagic lesions and yellow fluid). Blood was collected for enzyme investigation from the weekly slaughtered bird. Broilers had lower BWG and higher FCR when fed low EP feeds, regardless of the period fed when compared to the control (P < 0.001). Growth compensation, however, occurred afterwards such that all birds presented similar performance at the end. At 14, 21, and 28 d, broilers previously fed low EP feeds had lower WB scores (P < 0.001) compared to birds fed the control; however, both groups presented increased WB scores after 28 d. Wooden breast was positively correlated with breast echogenicity at 21 d (r = 0.31), 28 d (r = 0.43), 35 d (r = 0.21) and 42 d (r = 0.39). In conclusion, dietary energy and protein affected the development of WB scores in broilers and breast US images can be used as an early predictor of WB.