It has been reported that antipsychotics may improve cognitive function in the treatment of schizophrenia. The present study examined the effect of haloperidol and risperidone on cognitive performance in schizophrenic patients. 95 healthy subjects and 68 schizophrenic patients were recruited for comparison of cognitive function. As 20 of the 68 schizophrenic patients were drug-naive, they were randomly divided into two groups and double-blinded for treatment with either haloperidol or risperidone for an 8-week period. Each subject received Wisconsin Card Sorting Test (WCST) and Maze paradigms for cognitive function performance. For schizophrenic patients, the Positive and Negative Syndrome Scale (PANSS) was used for evaluation of clinical symptoms. Results demonstrated that in both WCST and Maze paradigms the 68 schizophrenic patients had worse cognitive performance compared with healthy subjects. Of the 20 drug-naive schizophrenic patients from the 68 in-patients, both haloperidol and risperidone improved the clinical symptoms. Maze tasks performance was improved progressively after haloperidol and risperidone treatment, although improvement was greatest with risperidone. Both haloperidol and risperidone had no evident effect on WCST performance. Our findings suggest that Maze paradigms may be an ideal tool for evaluation of pharmacological treatment effects on cognitive function in schizophrenic patients. Furthermore, risperidone may have more treatment benefits than haloperidol on cognitive performance in drug-naive schizophrenic patients.