ABSTRACT Background Childhood trauma is associated with increased rates of psychosis, with research identifying insecure attachment, dissociation and negative schemas as mediating factors. Network analysis offers a framework to explore complex interrelationships between symptoms and vulnerability factors. This study used network analysis to examine connections between experiences of childhood trauma, insecure attachment, dissociation, negative schemas and positive and negative psychosis symptoms. The study aims were to identify central, strongly connected variables and to compare network structures between subgroups with ( n = 259) and without ( n = 350) a self‐reported psychosis diagnosis. Methods Cross‐sectional, self‐report data from an online survey sample ( N = 609) were analysed. The sample was predominantly male (72.5%), White British/Other (80.1%), with a mean age of 28.70. The psychosis diagnosis sample contained more females (30.1%) than the sample without a psychosis diagnosis (21.1%). The psychosis diagnosis sample was also older, with a mean age of 32.93 compared with the no psychosis diagnosis sample (25.56). Pairwise Markov random field (PMRF) networks and strength centrality indices were used to identify central variables, while a network comparison test (NCT) evaluated structural differences between subgroups. Results Bizarre experiences showed strong connectivity with other positive symptoms. Negative‐self schemas and paranoia demonstrated strong cross‐construct connectivity, forming a cluster with negative‐other schemas, avolition and insecure attachment. Depersonalisation–derealisation was strongly connected to hallucinations. The NCT revealed no significant structural differences between subgroups. Conclusions Targeting negative‐self schemas and paranoia may yield the most substantial reductions in cross‐construct network connectivity and are indicated for treatment prioritisation. Interventions targeting detachment‐related dissociation may help reduce hallucinations. Longitudinal studies are needed to replicate these findings and explore causal pathways within network structures.