Gene drives are genetic systems designed to efficiently spread a modification through a population. They have been designed almost exclusively in eukaryotic species, especially in insects. We recently developed a CRISPR-based gene drive system in herpesviruses that relies on similar mechanisms and could efficiently spread into a population of wild-type viruses. A common consequence of gene drives in insects is the appearance and selection of drive-resistant sequences that are no longer recognized by CRISPR-Cas9. In this study, we analyzed in cell culture experiments the evolution of resistance in a viral gene drive against human cytomegalovirus. We report that after an initial invasion of the wild-type population, a drive-resistant population is positively selected over time and outcompetes gene drive viruses. However, we show that targeting evolutionarily conserved sequences ensures that drive-resistant viruses acquire long-lasting mutations and are durably attenuated. As a consequence, and even though engineered viruses do not stably persist in the viral population, remaining viruses have a replication defect, leading to a long-term reduction of viral levels. This marks an important step toward developing effective gene drives in herpesviruses, especially for therapeutic applications. IMPORTANCE The use of defective viruses that interfere with the replication of their infectious parent after coinfecting the same cells-a therapeutic strategy known as viral interference-has recently generated a lot of interest. The CRISPR-based system that we recently reported for herpesviruses represents a novel interfering strategy that causes the conversion of wild-type viruses into new recombinant viruses and drives the native viral population to extinction. In this study, we analyzed how targeted viruses evolved resistance against the technology. Through numerical simulations and cell culture experiments with human cytomegalovirus, we showed that after the initial propagation, a resistant viral population is positively selected and outcompetes engineered viruses over time. We show, however, that targeting evolutionarily conserved sequences ensures that resistant viruses are mutated and attenuated, which leads to a long-term reduction of viral levels. This marks an important step toward the development of novel therapeutic strategies against herpesviruses.