This study describes a general framework for predicting the accuracy of Mendelian sampling terms when truncation selection is applied on best linear unbiased prediction (BLUP) estimated breeding values. A selection index approach is followed. The pseudo-BLUP index is extended to include terms related to the Mendelian sampling term. Predicted accuracies are compared with those obtained through stochastic computer simulation. Good predictions for the accuracy of the Mendelian sampling term were obtained both at selection time and at convergence of long-term contributions of selected candidates for a range of heritabilities and population structures. The prediction approach developed provides a key tool for obtaining predictions of genetic response from quadratic optimization that maximizes the rate of genetic progress while restricting the rate of inbreeding.