In order to investigate muscle activity during overhand baseball pitching, surface EMG analysis of muscles in the shoulder girdle and upper limb was performed. The subjects were two college baseball players. Surface EMG was recorded from 25 portions of 16 muscles: the long and short head of the biceps brachii, the coracobrachialis, the brachialis, the lateral, medial and long head of the triceps brachii, the anterior, middle and posterior deltoid, the sternocostalis of the pectoralis major, the latissimus dorsi, the upper, middle and lower trapezius, three portions of the serratus anterior, the pronator teres, the flexor carpi radialis, the flexor carpi ulnalis, the extensor carpi radialis, the extensor digitorum, the extensor carpi ulnalis, and the brachioradialis. After rectifying the surface EMGs, their linear envelopes were extracted with a digital low-pass filter. The duration, timing and similarity of these muscle activities before and after ball-release were analyzed quantitatively through auto-correlation and cross-correlation analysis of the envelopes. The biceps and the brachialis were activated in the cocking phase and follow-through phase, and played a role in preparing for acceleration and deceleration in each phase. The triceps was activated in the acceleration phase and elbow joint extension, and contributed to the increase of ball speed by minimizing the moment of inertia about the longitudinal axis of the upper limb. The coracobrachialis was activated from the cocking phase until ball-release, and contributed to the horizontal extension of the shoulder joint. The latissimus dorsi and the pectoralis major were activated in the acceleration phase, and the duration of their activity was shorter than that of the other muscles. The activities of the trapezius and the serratus anterior differed between the two subjects. These activities in one subject were simultaneous in the acceleration phase, and those in the other were separate in the cocking phase. These activities probably represent differences in the control of scapula motion during pitching. The peak activities of the forearm muscles were concentrated in the ball-release phase, and the duration of activity was longer in the extensor muscles than in the flexor muscles. This type of correlation analysis is useful for extracting information about muscle activity during baseball pitching.