Abstarct: Computer models contain many inputs, the real value of their are uncertain. In current methods of sensitivity analysis for assessing the effects of the input uncertainties on the output of the computer models are used from Monte Carlo methods. For computer model with complex and expensive computational, it is often not possible to get a large enough sample for a meaningful uncertainty analysis. In this thesis, we are doing sensitivity analysis with presentation kriging model as a surrogate model, that is estimator of computer model, baxsed on the use of gaussian stochastic process models, in a Bayesian context. All of the results and measures provided by the Monte Carlo sensitivity analysis of the computer model are obtained with surrogate model, but using a far smaller sample. The accuracy of the kriging model is shown with an analytical function and a case study.