For personalized healthcare management, personal ECG records can help our system more precisely classify those arrhythmias which were diagnosed and detected before. The following results can prove that the SVM-based algorithm (RBF SVM) is effective and efficient because it has good accuracy performance and takes execution time less than NN method. Table 5 shows classification accuracy of multiple RRIs, used as attributes of each sample. The results present the classification accuracy of from 9-RRI to 72-RRI is better than 2-RRI. The detailed classification results of R207 by RBF SVM based on 2-RRI, 9-RRI, 18- RRI, 36-RRI and 72-RRI attributes are listed in Table 6.