A real time Electrocardiogram wave peak detection algorithm with low latency has been developed. It can detect different component wave peaks, viz., P, Q, R, S and T from digitized samples. The algorithm has been implemented using Xilinx Spartan-III xc3s400tq144-5 FPGA module. For the purpose of detection of ECG wave peaks, during a training period of 6 second, a slope and polarity based signature is learned from the incoming ECG signal, which is sampled at 1 kHz frequency. These detected signatures of individual wave peaks of the concerned are then used for detection of subsequent individual component wave peaks. R, T and P waves are detected with 98.8%, 97.98% and 97.75% respectively using ptbdb and mitdb data from Physionet. The design consumed 35% resources of FPGA. © 2014 IEEE.