T The heart, as a vital organ, has an important role in circulating blood throughout the body, but it is very susceptible to disease. To study and check the condition of the heart, an electrocardiograph (ECG) is periodically used, but noise or interference in the ECG signal retrieval often damages the accuracy of the data obtained. To overcome this, a 3-electrode ECG system based on an exponential moving average (EMA) filter is designed. The goal is to reduce the noise of the EKG signal and display the heart beats per minute on the computer. This system consists of 3 electrode sensors, the AD8232 module as an amplifier and ECG signal conditioning, NodeMCU as a data processor, and the Arduino IDE software as an application for uploading programs to NodeMCU. Tests in this study were carried out by comparing the results of ECG readings based on the EMA filter and ECG PC-80B against standard ECGs in hospitals. The results show that this system is successful in overcoming the problem of noise in ECG waves, making graphs smoother and easier to interpret. Furthermore, the system generates ECG signals in a format that is closer to hospital standard ECG than the PC-80B EKG device. This is evidenced by the QRS wave height in this system is not far from the EKG results at the hospital, with a difference of only 0.1 mV and a difference in the R-R wave interval of 6%. In addition, a comparison of the BPM results on this system was also carried out with the results of manual BPM testing without tools and showed a difference of 5%. This proves that the EKG system with an EMA filter has a pretty good level of accuracy