Objective: To explore the influencing factors and pathways of suicidal ideation among children and adolescents with severe autism spectrum disorder (ASD), so as to provide references for clarifying the impact intensity and pathways of various factors on suicidal ideation in the population. Methods: A cross sectional study was conducted from June 17, 2024, to January 12, 2025, involving 96 severely affected ASD children and adolescents aged 8-18 years from Tianjin. Participants were assessed using the Puberty Development Scale (PDS), Children s Alexithymia Measure (CAM), Strengths and Difficulties Questionnaire (SDQ), and Positive and Negative Suicide Ideation (PANSI). The random forest Boruta algorithm was employed to screen core variables, and a Bayesian network model was constructed to analyze the influencing factors of suicidal ideation in children and adolescents with severe ASD. Results: Through the screening using the Boruta algorithm, the SDQ scale score, conduct problems, hyperactivity, peer relationship problems and prosocial behavior were identified as the key predictors of suicidal ideation. A Bayesian network model was established with hyperactivity as the central mediating node. The impact of hyperactivity on suicidal ideation exhibited a non linear relationship: compared to the normal state (31.6%, 68.4%), the borderline state of hyperactivity was associated with a higher probability of low risk suicidal ideation (47.1%) and a lower probability of high risk suicidal ideation (52.9%). Suicidal ideation among children and adolescents with severe ASD was closely related to hyperactivity. In the state of hyperactivity, the abnormal peer relationship (95.2%) and the abnormal prosocial behavior (77.0%) were aggravated. Conclusions Suicide ideation among children and adolescents with severe ASD is strongly associated with hyperactivity traits. It is necessary to establish a prevention and control system centered on hyperactivity intervention to reduce this risk.