Abstarct: Heart and blood transfer system are the most important organs in the human body. Every disruption in the function of these organs can have a major effect on body because the blood transfer system supplies the requisite energy of all body organs (even the heart). Heart diseases are the main cause of death which kill or hurt a lot of people each year. The number of death caused by heart diseases is greater than any other natural incident. This fact has contributed to extended developments in medical and heart diseases also some engineering sciences are applied to prevent the occurrence of these diseases. This paper tries to extend the detection algorithms of heart Arrhythmias using data heart measurement systems like Electrocardiogram signal. In order to extract changes of heart signals regarding their figures, frequency and time features of the heart signals are inspected via different methods and finally Wavelet method is selected. db4 is selected among Wavelet functions and values of partial signals along the energy signal are selected after applying this Wavelet for 5 steps. After extracting features, in order to decrease dimension of feature space, PCA operator is used and the length of feature vector drops to 15. Attempt has been made to include design with faster training approach, selection of optimal parameters, higher authenticity and very low computation volume, in order to design different classifiers (such as Bayesian, k-nearest neighbor, Parzen and multilxayer neural network). The precision of basic classifiers for 5 proposed classes is between 85 to 90 percent which can be raised to 100 percent using proposed algorithms like OWA, fuzzy integrals, Dempster-Shafer and Mixture of experts. It can be applied in this field for online implementation (considering low computation volume) with features of the presented model.