This study sought to develop and validate a screening battery for detecting HIV-related neuropsychological (NP) impairment. Six NP measures representing the ability areas most likely affected by HIV infection were paired in 14 combinations and their diagnostic accuracy rates compared. The measures were selected from a larger NP battery administered to 190 HIV-seropositive (HIV+) participants. Screening battery performance was classified as NP impaired if demographically corrected T-scores fell below 40 on both tests, or below 35 on one test. Using blind clinical ratings of NP test results from the larger battery as the "gold standard" for global NP status (impaired or unimpaired), we found that several test combinations demonstrated adequate diagnostic accuracy in detecting NP impairment. The most sensitive test combinations were the Hopkins Verbal Learning Test-Revised (HVLT-R; Total Recall) and the Grooved Pegboard Test nondominant hand (PND) pair and the HVLT-R and WAIS-III Digit Symbol (DS) subtest pair (sensitivity = 78% and 75%, respectively). Both test combinations (HVLT-R/PND, HVLT-R/DS) were more accurate than the HIV Dementia Scale (HDS) in classifying HIV+ participants as NP impaired or unimpaired. Results suggest that demographically corrected T-scores from pairs of common NP measures may serve as valid screening instruments to identify subjects with HIV-related neurocognitive impairment who could benefit from more extensive NP examination.