Ventricular Tachycardia (VT) and Ventricular Fibrillation (VF) are fatal cardiac diseases associated with cardiac arrest. It is difficult to manually classify VT and VF signals. However, precise classification of VT and VF signals can assist cardiologists to identify and ultimately prevent onset of VF or VT. In this thesis, some of the underlying features which characterize VF and VT are extracted and are used to efficiently classifying these signals. The features are acquired from energy coefficients matrices using Continuous Wavelet Transform (CWT) through application of Principal Component Analysis (PCA). The features are the vector containing newly generated energy projection coefficients and the vector containing the number of the top 99% principal components (Eigen-Values) for each case. Feature vectors are then passed through Fast Forward Neural Network (FFNN) and Leave One Out Method (LOOM) classifiers for discrimination. The results are then compared for the highest classification results for VF and VT signals.