Borderline personality disorder (BPD) has been associated both with frontolimbic dysfunction and maladaptive personality traits such as impulsivity and neuroticism. The purpose of this study was to examine whether there are differences in predictive variance between neuropsychological variables and selected personality traits in persons with BPD. We compared persons with DSM-IV BPD (n = 25) to individuals without BPD or substance abuse (n = 20). All subjects completed a battery of neuropsychological tests, Cloninger's Temperament and Character Inventory, and the Barratt Impulsiveness Scale. BPD subjects had evidence of cognitive inhibition, and defi cits in working memory, perseveration, and decision-making not accounted for by IQ differences between groups. In particular, Stroop Interference remained signifi cant after controlling for IQ, depression and alcohol use (p = 0.043). They also had higher levels of impulsivity, novelty seeking and harm avoidance, but lower levels of self-directedness and cooperativeness than the comparison group. Logistic regression analyses indicated that both personality traits (Nagelkerke R 2 = 0.74, Model χ 2 = 36.21, p < 0.001) and neuropsychological tests (Nagelkerke R 2 = 0. 47, Model χ 2 = 19.22, p < 0.001) were signifi cant predictors of BPD. However, neuropsychological tests did not add signifi cant incremental variance over selected personality trait dimensions in predicting BPD (Block χ 2 = 6.45, p = 0.168), whereas personality traits added signifi cant incremental variance over neuropsychological tests in predicting BPD (Block χ 2 = 3.44, p < 0.001). This study suggests that persons with BPD have impaired neuropsychological performance, but that impulsivity and other selected personality traits have a primary role in predicting BPD over and above neuropsychological test abnormalities.