Abstract: Objective To evaluate the re-bleeding risk severity and the clinical efficacy of clinical grading nursing intervention in patients with acute non-variceal upper gastrointestinal bleeding (ANVUGIB) using Rockall risk scoring.Methods A total of 56 patients from January to December 2014 were selected as control group while another 51 cases from January to December 2015 were as experimental group.Control group was added with routine nursing while experimental group with clinical grading nursing intervention based on Rockall risk scoring on the basis of treatment for gastrointestinal bleeding.The occurrence rate of re-bleeding during hospitalization was observed inboth groups,the improvement of anxiety and depression as well as patients'cognition on disease and behaviors following doctors'advice were compared between two groups.Results The total occurrence rate of re-bleeding was 7.8％ in experimental group,evidently lower than the 25.0％ in control group (P ＜0.05).After intervention,self-rating depression scale (SDS) and self-rating anxiety scale (SAS) decreased in both groups (P ＜0.05 or P ＜0.01),which decreased more significantly in experimental group than that in control group (P ＜0.01).Experimental group was markedly higher in the number of patients who grasped the knowledge and cognitive behaviors on disease (P ＜ 0.05 or P ＜ 0.01) as well as the compliance of behaviors following doctors'advice than control group (P ＜ 0.05).Conclusion Clinical grading nursing intervention based on Rockall risk scoring can effectively reduce the re-bleeding risk of patients with ANVUGIB,relieve their anxiety and depression,and improve their cognition on disease and behaviors following doctors'advice.