Estimations of shoulder muscle load are important in biomechanic and ergonomic research. We have studied shoulder muscle load in the trapezius (six subjects), deltoid (six subjects), infraspinatus (eight subjects), and supraspinatus (seven subjects) muscles with simultaneous intramuscular pressure (IMP) and intramuscular bipolar electromyography (EMG) recordings. For imposition of shoulder muscle load, the arm was positioned in abduction or flexion with different hand loads (0, 1, or 2 kg), or isometric force registrations were performed. The microcapillary infusion technique was used for IMP recordings. The IMP in the supra- and infraspinatus muscles were high compared with the trapezius and deltoid muscles in abducted arm positions. In all test situations, IMP and EMG gave a similar description of local muscle load. IMP at maximal voluntary contraction was highest in the supraspinatus and infraspinatus muscles. Both IMP and EMG in all four muscles showed an almost linear correlation to recorded isometric external force. The difference in IMP between shoulder muscles in the same arm position may be due to muscle anatomy, muscle function, and compliance of surrounding tissues. Because a high IMP may impede muscle blood flow, our findings may possibly explain the physiological stress on the rotator cuff muscles as compared with the deltoid and trapezius muscles in work with elevated arms.