Objective : To analyze and verify the factors influencing the prediction model of suicidal ideation of burn patients in hospital based on international scale． Methods : The clinical data of 194 burn patients treated in hospi- tal were retrospectively analyzed．General data questionnaire，ISI，HAMD，HAMA，ASDS and BSHS-B were used to evaluate the influencing factors of suicidal ideation．According to the presence or absence of suicidal ideation， the patients were divided into the suicidal ideation group and the non-suicidal ideation group．The baseline data be- tween the groups were compared，univariate screening of meaningful variables was conducted，and multivariate Lo- gistic regression modeling was further conducted. ROC analysis evaluated model differentiation，and internal verifi- cation was conducted. Results : According to the baseline data analysis results，there were no statistically signifi- cant differences in age，BMI，years of education，smoking history，estimated percentage of burned area，head and neck burns，hip and perineal burns，and pain scores in the suicidal ideation group (21 /194) compared with the non-suicidal ideation group ( 173 /194) ．Gender (P = 0. 047) ，presence or absence of trunk burn (P = 0. 022) ， severity of burn ( moderate burn : P = 0. 002 ; severe burn : P = 0. 458 ; extremely severe burn : P = 0. 169) ，ISI score (P = 0. 001) ，HAMD score (P = 0. 001) ，HAMA score (P＜0. 001) ，ASDS score (P = 0. 003) ，BSHS-B score (P = 0. 011) had statistical significance．Multivariate Logistic regression analysis showed that the severity of burn (moderate burn : OR = 0. 103，P = 0. 009 ; severe burn : OR = 0. 351，P = 0. 223 ; extremely severe burn : OR = 0. 103，P = 0. 095) and HAMA score ( OR = 1. 136，P = 0. 007) were independent influencing factors for burn patients with suicidal ideation 。The Logistic regression prediction model was established by two independent influ- encing factors. ROC analysis results showed that the model had good differentiation ( AUC = 0. 880，95% CI : 0. 808 －0. 952 ，P ＜0. 001 ) and the internal verification accuracy was 79. 38%． Conclusion The prediction model built on the basis of two independent influencing factors，burn severity and HAMA score，has a good predic- tion accuracy，which is helpful for clinicians to intervene as soon as possible for burn patients with suicidal ideation in hospital，in order to reduce the incidence and enrich clinical psychological research.