Objective To develop and validate a prediction model for severe disability or death(SDD)in acute ischemic stroke(AIS)patients who underwent endovascular treatment(EVT).Methods Based on the dataset of ANGEL-ACT study who received EVT for AIS between november 2017 and march 2019,a retrospective analysis was performed on 1 677 patients,including 1 111 males and 566 females,aged ≥ 18 years.Patients were divided into two groups according to whether SDD occurred(mRS 5-6 scores 90 days after surgery):SDD group(n=478)and non-SDD group(n=1 199).Risk factors that might influence SDD after EVT in AIS patients were screened and analyzed by multifactorial analysis,LAS-SO regression,and RF-RFE methods.A nomogram was developed after evaluating the model performance and the execution of internal validation.Results SDD occurred in 380(28.1％)patients in the develop-ment cohort and 98(30.2％)patients in the validation cohort.Combining the three variable screening meth-ods,10 risk factors were selected for inclusion in the final model:age,NIHSS score,whether successful re-canalization,glucose level,hemoglobin,hematocrit,onset to puncture time,systolic blood pressure,AS-PECT score,and whether have treatment-related serious adverse events.A two-stage model means that model 1 contains pre-treatment variables(7 in total)and model 2 contains pre-treatment and post-treatment variables(10 in total).The area under the curve(AUC)of model 1 in the development cohort was 0.705(95％CI 0.674-0.736)and 0.731(95％CI 0.701-0.760)in model 2.Both models had good calibration with aslope of 1.000,and the decision curve analysis showed good clinical applicability.The results of the validation cohort were similar to those of the development cohort.Conclusion Age,admission NIHSS score,whether successful recanalization,admission glucose level,hemoglobin content,erythrocyte pressure volume,onset to puncture time,admission systolic blood pressure,ASPECT score,and whether have treat-ment-related serious adverse events are risk factors for SDD in patients with acute ischemic stroke.The two prediction models based on the above factors were used before and after endovascular treatment to predict SDD occurrence better.