The objectives of this work were to determine optimal surface electromyogram (EMG) electrode locations, and inter-electrode distance (IED), when assessing activity and fatigue in the human upper trapezius muscle. Surface EMG signals were recorded from the upper trapezius muscle of 11 healthy male subjects using a linear array of 16 surface electrodes. Five arm positions were investigated (arms at the side of the body, 45 degrees and 90 degrees flexion, 45 degrees and 90 degrees abduction). Fatiguing (1 kg hand load held for 3 min) and non-fatiguing (no load, 0.5 kg and 1 kg hand load held for 3 s) contractions were made. The variabilities of the average rectified value, root mean square, mean and median power spectral frequency and slope over time of these parameters as functions of electrode location and IED (from 5 mm over a range of 35 mm in steps of 5 mm) were quantitatively evaluated. A criterion for selecting the optimal electrode position was applied. This criterion indicated an optimal location measured from the acromion (38% of the distance from the lateral edge of acromion to the spine of the seventh cervical vertebra) which was statistically the same for all the EMG descriptors, arm positions and IED investigated. Finally, it was found that both EMG variables and indexes of muscle fatigue depended on IED which should thus be properly standardised. On the basis of the sensitivity of the EMG descriptors to electrode location and cross-talk reduction, an IED of 20 mm is suggested when a global analysis of activity in the upper trapezius muscle is made using a single pair of electrodes. This study emphasises that a surface EMG analysis of the upper trapezius muscle, following a proper placement of the electrodes and selection of IED, can give reliable indications of muscle activity and fatigue. Data on the myoelectric manifestations of muscle fatigue of the upper trapezius muscle are provided for the optimal electrode location.