Recording of the electrical activity associated to heart functioning is known as Electrocardiogram (ECG). ECG is a quasi-periodical, rhythmically signal synchronized by the function of the heart, which acts as a generator of bioelectric events. ECG signals are low level signals and sensitive to external contaminations. Electrocardiogram signals are often corrupted by noise which may have electrical or electrophysiological origin. The noise signal tends to alter the signal morphology, thereby hindering the correct diagnosis. In order to remove the unwanted noise, a digital filtering technique based on adjustable windows is proposed in this paper. Finite Impulse Response (FIR) low pass is designed using windowing method for the ECG signal. The results obtained from different techniques are compared on the basis of popularly used signal error measures like SNR, PRD, PRD1, and MSE.