This work describes the development of a technique for the evaluation of the performance of a product's physical user interface. The technique is intended to combine the best features cat conventional user group testing with those of computer based biomechanical modelling. A requirement for the new technique exists as social pressure demands that consumer products he optimised for users with a wide range of physical capabilities, while shortening product lifecycles leave less time for extensive user evaluation programmes. A demonstration system was developed, based upon the use of an electromagnetic tracking system to gather upper limb motion data and a two segment, rigid link biomechanical model. Experimental work was carried out to test the effectiveness of the system at following limb movements and average error in reconstruction of hand position from segment angle data was 62mm (Standard deviation 41 mm) The modelling system was applied to the assessment of two types of product: cutlery and drinking vessels and the effectiveness of various statistical techniques in allowing the rapid identification of important design parameters was assessed. The use of Taguchi's smaller-the-better signal to noise ratio was found to be effective for the measurement of the effect of product design on shoulder and elbow forces. Cutlery with enlarged handles designed to reduce grip strength requirements tended to increase forces at the shoulder. The method was also applied to an interface optimisation problem involving the design of a lever mechanism. Partial factorial design was used to minimise experimental cost during the assessment of multiple factors, but strong interactions were detected between interface parameters, reducing the value of the analysis. The overall height of the lever handle relative to the user's shoulder was found to be the most significant design factor, with an optimum operating situation existing where the lever was low enough to require almost full extension of the elbow during use. The work concludes that biomechanical analysis holds further promise for the optimisation of interface parameters, provided the high experimental cost involved with present techniques can be reduced.