The research work presented in my thesis aims to manage geometrical variability throughout product life cycle and its consequence on the product performance. The geometrical variations generated from the manufacturing to assembly stage are modeled by the geometrical deviation model. Monte-Carlo simulation method is then used to generate an image of the real manufactured product. As a result, the geometrical deviations are integrated into simulation of product performance in order to establish the relationship between the performance and the parameters of geometrical deviations or variation sources. An image of real performance of the manufactured product is then generated. From this result, the parameters of variation sources influencing the product performance are identified and classified according to their impact level. The variance of the product performance variation is also established by the relation between the performance and the parameters of geometrical deviations or variation sources. Finally, the robust design solution can be found by minimization of the variance of the product performance variation.