Cardiovascular disease has always been a major factor threatening human life and health. If the pathological information contained in human heart sound signals can be accurately classified， it will be very helpful for disease diagnosis and control. Firstly， particle swarm optimization algorithm is used to optimize the traditional support vector machine algorithm， and a binary classifier model is proposed. The primary classifier is composed of three algorithms Adaboost， RF and PSOA－SVM based on Stacking method， and the secondary classifier is LR model； Secondly， the improved Grey Wolf Optimization Algorithm is used to find the optimal parameter combination of support vector machine to get a new classifier model； Finally， the heart sound data set is used to analyze the two classifier models. The experiments show that the two models show excellent classification results.