Neurological soft signs (NSS) are one of the biomarkers for schizophrenia spectrum disorders. However, a few studies have examined the prevalence of NSS across the schizophrenia spectrum. The present study adopted a quasi-longitudinal study design and examined the prevalence of NSS and their associations with clinical and behavioural manifestations in participants in different stages of the illness. The abridged version of the Cambridge Neurological Inventory was administered to 39 patients with the first-episode schizophrenia, 39 individuals with ultra-high risk (UHR) for psychosis, 39 individuals with schizotypy, and 39 healthy controls. Patients with the first-episode schizophrenia had a higher prevalence of NSS in motor coordination than healthy controls as well as individuals with UHR and schizotypy. Individuals with UHR exhibited a higher prevalence of sensory integration items than individuals with schizotypy and healthy controls. Discriminant analysis classified the membership of the individuals correctly across the spectrum with an accuracy of up to 60.9%. In particular, NSS could discriminate individuals with UHR from healthy controls at up to 85.9% accuracy. These findings suggest that NSS are robust biomarkers to detect and discriminate individuals in different stages of the schizophrenia spectrum from healthy controls.