Patients with schizophrenia have consistently been found to exhibit cognitive deficits, particularly in memory, which have been suggested to mediate functional outcomes. Several recent reviews of cognitive retraining have concluded that these deficits respond to training, although the sustainability of cognitive improvement following completion of training has not been adequately evaluated. Most studies had small samples and very short follow-up periods. As part of a larger study, we examined performance on two memory tasks in two groups of participants: those who received computerized cognitive remediation training in addition to work therapy (n=45), vs. those who only received work therapy (n=49). Computerized cognitive remediation included hierarchical training on a computerized digit span task and a computerized words serial position task. Assessments using the same computerized tasks were made at three time points: baseline, end of active intervention, and 6-month follow-up. Compared to the active control condition (work therapy only), the group receiving computerized cognitive remediation in addition to work therapy showed significantly greater improvements on the trained digit span task following training. These improvements were maintained at the 6-month follow-up. There were no significant group differences on the word serial position task at any time point. Results indicate that computerized training can lead to sustained improvements on some, but not all, training tasks.