For the models with interval input variables and random input variables simultaneously, a validation metric was researched to measure the agreement between the quantitative predictions from an uncertain model and relevant empirical data. With respect to different sources of uncertainty existed in engineering mathematic models and experiments，the characteristic of the validation for models with stochastic and interval mixed uncertainty was analyzed. A new validation metric for model with stochastic and interval mixed uncertainty was proposed by using the interval theory and the probability method. The properties of the proposed validation metric were discussed, and its calculation method and procedures were presented. A numerical test case and an engineering example were used to verify the feasibility and effectiveness of the proposed validation metric.