Objective To screen fall risk in community elderly population by employing the fall risk self-assessment scale of the STEADI (Stopping Elderly Accidents, Deaths, & Injuries), and to test the predictive validity for long-term fall accidents. Methods Multi-stage cluster random sampling method was used to select 1 168 community elderly persons in Chongqing, and then their fall risk was assessed using the Chinese version of the fall risk self-assessment scale of the STEADI.Fall accidents were recorded at 6-month follow-up by telephone. Results Of the 1 168 participants, 321 （27.5%） reported fall risk and 825 were followed up. Sixty-four participants reported fall accidents at 6-month follow-up, with the prevalence being 7.8%, and 46 (71.9%) of them were injured. The sensitivity and specificity of the Chinese version of the scale for predicting the occurrence of falls after 6 months were 45.3 % and 75.4 % respectively.The AUC was 0.604 and the Youden index was 0.208. More than seventy percent (71.6%) of the elderly thought that self-assessment of fall risk was meaningful, and 69.9% of them had implemented measures to prevent falls after the assessment. Conclusion Majority of the elderly recognize the significance of self-assessment of fall risk and take preventive measures.However, the Chinese version of the STEADI self-assessment scale is not so satisfactory to predict falls after 6 months in the community elderly,and it is recommended to combine with gait, strength and balance results for fall risk prediction.