Summary: The missing response problem in nonlinear semiparametric EV (error-in-variables) models is considered, where the explanatory variate is measured with error. With the help of validation data, two estimators for the unknown parameters and the nonparametric function in the model are proposed. The asymptotic normality of the proposed statistics for the unknown parameters as well as the optimal convergence rate for the statistics of the nonparametric function is obtained.