Aim There is an increasing need for identifying neurocognitive predictors of global functional outcome in early psychosis toward optimizing an early intervention strategy. Methods We conducted a longitudinal observational study to investigate an association between neurocognitive assessments at baseline and global functional outcome at an average of 1‐year follow up. Participants included ultra‐high‐risk for psychosis (UHR) individuals who had not converted to psychosis during the follow‐up period (UHR‐NP) and those with first‐episode psychosis (FEP). We evaluated neurocognition at baseline using the Brief Assessment of Cognition in Schizophrenia Japanese version, including Verbal Memory, Working Memory, Motor Speed, Verbal Fluency, Attention/Processing Speed, and Executive Function. We also assessed global functional outcome using the modified Global Assessment of Functioning (mGAF) scale both at baseline and after the follow‐up period. Results Thirty‐four UHR‐NP individuals (34/47, 72%) and 29 FEP individuals (29/36, 81%) completed assessment of neurocognitive function at baseline and functional outcome at follow up. In the UHR‐NP group, Attention/Processing Speed was significantly associated with the mGAF score at follow up. In the FEP group, Executive Function was significantly associated with the average mGAF score during follow up. Conclusion Attention/Processing Speed and Executive Function at baseline may predict global functional outcome of early psychosis. These neurocognitive tests are easy to incorporate in clinical settings and, if replicated in independent samples, may be included in routine clinical assessments for prediction of functional outcome in early psychosis.