OBJECTIVE: Previous studies suggested that schizotypal traits are closely related with affective symptoms and social dysfunction in general population. However, the dynamic interactions among these variables across time remain unclear. This study aimed to investigate this issue using cross-sectional and longitudinal network analysis. METHODS: A total of 9136 college students completed self-report scales measuring schizotypal traits, anxiety, depression, prodromal risk, and social functioning at timepoint 1 (T1) and six months later. Classical contemporaneous network analysis was used to estimate the associations among all variables at T1, and the cross-lagged panel network analysis was performed to estimate the predictive effects at follow-up. In addition, prodromal high-risk and low-risk subgroups at T1 were identified using Prodromal Questionnaire (PQ-16) to investigate differences on network constructs and edge weights using Network Comparison Tests (NCT). RESULTS: Contemporaneous network showed that negative dimension of schizotypal traits, anxiety, and depression interacted and were all negatively correlated with social functioning. Of all the nodes in the network, interpersonal features of schizotypal traits was the most centralized. The results of cross-lagged network analysis further verified the predictive effect of personality traits and emotions on social function. In addition, NCT revealed significant difference between prodromal high-risk and low-risk subgroups in network structure and strength of edges connecting schizotypal traits, affective symptoms with prodromal risk, but not in global strength. CONCLUSIONS: Our findings suggest that schizotypal traits, affective symptoms, and prodromal risk relate to social functioning, both independently and through their dynamic interactions.