Unstable angina(UA) is the most dangerous type of Coronary Heart Disease(CHD) to cause more and more mortal and morbid world wide. Identification of biomarkers for UA at the level of proteomics and metabolomics is a better avenue to understand the inner mechanism of it. Feature selection based data mining method is better suited to identify biomarkers of UA. In this study, we carried out clinical epidemiology to collect plasmas of UA in-patients and controls. Proteomics and metabolomics data were obtained via two-dimensional difference gel electrophoresis and gas chromatography techniques. We presented a novel computational strategy to select biomarkers as few as possible for UA in the two groups of data. Firstly, decision tree was used to select biomarkers for UA and 3-fold cross validation was used to evaluate computational performances for the three methods. Alternatively, we combined independent t test and classification based data mining method as well as backward elimination technique to select, as few as possible, protein and metabolite biomarkers with best classification performances. By the method, we selected 6 proteins and 5 metabolites for UA. The novel method presented here provides a better insight into the pathology of a disease. Unstable angina (UA) is the most dangerous type of Coronary Heart Disease (CHD) to cause more and more mortal and morbid world wide. Identification of biomarkers for UA at the level of proteomics and metabolomics is a better avenue to understand the inner mechanism of it. Feature selection based data mining method is better suited to identify biomarkers of UA. In this study, we carried out clinical epidemiology to collect plasmas of UA in-patients and controls. Proteomics and metabolomics data were obtained via two-dimensional difference gel electrophoresis First, decision tree was used to select biomarkers for UA and 3-fold cross validation was used to evaluate computational performances for the three methods. Alternatively, we combined independent t test and classification based data mining method as well as backward elimination tech nique to select, as few as possible, protein and metabolite biomarkers with best classification performances. By the method, we selected 6 proteins and 5 metabolites for UA. The novel method presented here provides a better insight into the pathology of a disease.