Bus reliability is essential for maintaining the steady state of electrical network in power plants and power transformer stations. Researches on techniques and methods to achieve highly reliable and intelligent bus protection are very important. A new bus protection method based on parameter optimization of the support vector machine (SVM) is provided in this study. Various types of fault data were collected and used as samples for sorting comparison and analysis of different bus faults. Data were obtained by comparing the SVM method with the artificial neural network (ANN) method. The results suggest that the SVM model using RBF kernel can better distinguish among bus normal operation and different bus fault types than the ANN model, and meet the precision requirements for bus protection.