Abstract: The modern modelvalidation considers uncertainty and the modeling error. In order to study modeling and validation criterions in the model validation course, in this paper, the normaldistribution and nonparametric kernel density estimate (KDE) are adopted totest the uncertain modeling by an example of the Sandia's structuraldynamics challenge problem. At the same time, two validation criterions areadopted for model validation to obtain a better model.It is shown that by combining twovalidation criterions in the processes of model validation, betterresults can be expected.