Frequentist model average estimation receives much attention in recent years.However, no investigation has been conducted for the data with measurement errors. In this paper, we consider frequentist model average estimation for linear errors-in-variables models.The asymptotic distribution of the model average estimator is derived, and a confidence interval having a coverage probability that tends toward the nominal level in large samples is constructed. Further, the confidence interval constructed based on the model average estimator is shown to be asymptotically the same as that obtained under the {\rm full} model. A
simulation study shows that the finite sample performance of the model average estimator is better than that of the model selection approach or of the {\rm full} model approach.