This document presents a method for compressing ECG signal data using neural networks. Twelve features are extracted from echocardiogram data and used as input to a neural network with a dual three-layer backpropagation structure. The network is trained and tested on a dataset using backpropagation algorithm, achieving 99.5% efficiency. Backpropagation is used to adjust the weights in the neural network to map inputs to the correct outputs. The study demonstrates that neural networks can effectively compress ECG signal data for applications like telemedicine.