Abstract There is considerable evidence that cognitive impairment is a better predictor of work and social function in schizophrenia than are positive and negative symptoms. Atypical antipsychotic drugs have been shown to improve cognitive function in schizophrenia patients, but it is unclear whether this improves patients' ability to gain employment. Data from a prospective longitudinal study was used to test the hypotheses that (1) clozapine treatment would improve employment outcome in treatment-resistant schizophrenia or schizoaffective disorder patients, and (2) specific cognitive functioning at baseline and after treatment would predict work status at baseline and change in work status. Employment status and cognitive assessment data were collected in 59 treatment-resistant schizophrenia or schizoaffective disorder patients. Forty-seven of 59 (79.7%) patients were unemployed at baseline. Over a 12-month period, 23 (48.9%) additional patients were able to gain paid or volunteer jobs, or attend school. As predicted, neurocognitive performance was a better predictor of employment status and ability to gain of employment than clinical symptoms. Improvement in verbal working memory was found to be a better predictor of employment outcome than other cognitive functions. Treatment that enhances cognitive function, especially verbal working memory, may lead to better employment outcomes in treatment-resistant schizophrenia or schizoaffective disorder patients.