Objective: The aim of the present study was to investigate and compare cognitive functioning of first-degree relatives of people with schizophrenia who were also at ultra-high risk (UHR) for psychosis with patients with first-episode (FE) schizophrenia, first degree relatives of patients not fulfilling UHR criteria (FDR), and healthy control (HC) subjects. Method: Forty subjects in each group were included, underwent a face-to-face interview and completed a neurocognitive test battery, including the Trail Making Test-A (TMT-A, psychomotor functions), Stroop Color Word Test (attention), Digit Symbol Coding Test (DST, processing speed and working memory) and Hopkins Verbal Leaning Test-Revised (HVLT-R, verbal memory). Results: Functioning in all the cognitive test domains displayed a gradual decrease from the HC, FDR, UHR to FE groups. After controlling for covariates, there were still significant differences in TMT-A (F = 35.4, P < 0.001), DST (F = 38.9, P < 0.001), Stroop Color Word Test (F = 35.0, P < 0.001), Stroop Word Test (F = 36.2, P < 0.001), Stroop Color Test (F = 40.9, P < 0.001) and HVLT-R (F = 62.5, P < 0.001) between the four groups, indicating that the cognitive functioning in the UHR group was intermediate between the FE and FDR groups, while the FDR group had poorer performance than the HC group, and the FE group had the poorest cognitive functioning across all four examined domains. Conclusion: The results indicate that impairments in processing speed, attention, working memory and verbal memory exist in both UHR and FDR subjects. In order to clarify the associations between cognitive functioning and UHR and schizophrenia, longitudinal studies are warranted.