Autism spectrum disorder and psychotic disorders, although clinically distinct, share overlapping characteristics, particularly in the domain of social cognition. Both autistic traits and psychotic-like experiences (PLEs) are increasingly conceptualized as existing along a continuum within the general population. Empathy, a fundamental aspect of social cognition, is commonly associated with both conditions. This study examined the relationships among autistic traits, PLEs, and empathy in a large general-population sample of preadolescents using network analysis. Data were drawn from the Adolescent Brain Cognitive Development Study, including 9,214 participants (age range = 8.92 to 11.08 years; 4,850 males). Autistic traits were assessed using the abbreviated version of the Social Responsiveness Scale, Second Edition, which measures social and communication difficulties and restricted and repetitive behaviors. PLEs were evaluated using the Prodromal Questionnaire, Brief Child Version, which assesses hallucinations, thought delusions, and grandiose delusions. Associations among these constructs were modeled using both an undirected Gaussian graphical model and a directed acyclic graph (DAG). Higher empathy scores were negatively associated with elevated autistic traits (social and communication interactions and restricted and repetitive behaviors) and with PLEs (grandiose delusions). Moreover, higher levels of autistic traits were positively associated with greater distress related to grandiose delusions and hallucinations. Centrality analysis identified hallucinations as a key node in the network, a result supported by the DAG. Sex-specific analyses revealed subtle differences in network connectivity between males and females. These findings highlight the intricate interplay among autistic traits, PLEs, and empathy during preadolescence and emphasize empathy’s negative relationship with autistic traits. Overall, the results offer insight into shared social cognitive processes across neurodevelopmental and psychosis-spectrum traits.