Electrocardiogram (ECG) can provide valuable clinical information on cardiac functions. This paper illustrates an algorithm for real-time detection of wave peaks and their features from single lead ECG data. At first, the ECG data was filtered for power line interference and high frequency noise. Then, a set of slope and polarity-based rule bases were generated from the first 6000 samples, which define templates of R-peak, P-and T-wave detection from the following beats. The algorithm was implemented on Xilinx Spartan III Field Programmable Gate Array (FPGA). For testing of the algorithm, ECG data was quantised at 8-bit resolution and delivered to the FPGA using synchronous transfer mechanism using parallel port of computer. Xilinx implementation results provided 97.58%, 98.4% and 97.78% detection sensitivity for P-, R- and T-waves, respectively. Different wave features (height, polarity and duration) were detected with an average error rate of 9.3%. The detected wave signatures were clinically validated by medical expert. Copyright © 2015 Inderscience Enterprises Ltd.