Background: Neurocognitive impairments experienced by individuals at ultra-high risk (UHR) for psychosis are potential predictors of outcome within this population, however there is inconsistency regarding the specific neurocognitive domains implicated. This study aimed to examine whether baseline neurocognition predicted transition to psychosis, or functional outcomes, at medium-term (mean = 3.4 years) follow-up, while controlling for other clinical/treatment variables associated with transition to psychosis. Method: Analysis of data collected as part of a multi-centre RCT of omega-3 fatty acids and cognitive-behavioural case management (NEURAPRO) for UHR individuals was conducted on the 294 participants (134 males, 160 females) who completed neurocognitive assessment (Brief Assessment of Cognition for Schizophrenia) at baseline. Transition to psychosis was determined using the Comprehensive Assessment of At-Risk Mental States (CAARMS), and functioning was measured with the Global Functioning: Social and Role Scales. Results: Mean baseline z-scores indicated that UHR participants performed a quarter to half a standard deviation below normative means in all domains (range mean z = −0.24 to −0.47), except for executive functioning (mean z = 0.16). After adjusting for covariates, poorer Executive (p = .010) and Motor (p = .030) functions were predictive of transition to psychosis. Processing Speed and Verbal Fluency were significant predictors of role functioning at 12 months (p = .004), and social functioning at medium-term follow-up (p = .015), respectively. Conclusions: Neurocognitive abilities are independent predictors of both transition to psychosis and functional outcomes within the UHR population. Further research is needed to determine the best combination of risk variables in UHR individuals for prediction of psychosis transition, functioning and other psychopathology outcomes.