Automatic extraction of time plane features is important for cardiac disease diagnosis. ECG signals commonly change their statistical property over time and are highly non- stationary signals. For the analysis of ECG signals wavelet transform is a powerful tool. This paper presents a discrete wavelet transform based system for detection and extraction of P wave, QRS complex, and ST segment. The features like amplitude, frequency, energy are extracted from the Electrocardiogram (ECG) to classify them into normal and arrhythmic. The extracted features are given as input to neural network to classify them into normal and arrhythmic. The algorithm was implemented in MATLAB and the same was implemented in real time using Lab VIEW by acquiring the signal from subjects using BioKit(3-lead ECG).The above wavelet technique provides less computational time and better accuracy for classification, analysis and characterization of normal and abnormal patterns of ECG.