Sudden Cardiac Arrest (SCA) is a major cause of death due to cardiovascular diseases, responsible for approximately 4–5 million fatalities every year worldwide. However, prompt medical intervention and proper treatment can significantly improve survival rates for individuals experiencing SCA. In this work, we propose a hybrid deep learning model that fuses 1D-Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks for the early prediction of SCA using Electrocardiogram (ECG) signals. The 1D-CNN component is used to capture spatial features and irregularities associated with SCA, while the LSTM layer identifies temporal patterns essential for accurate prediction. The proposed model achieves a training accuracy of 99.93% and a testing accuracy of 99.75% predicting SCA from ECG signals 30 min prior to its occurrence. For signals recorded 60 min before SCA onset, the model reports a training accuracy of 99.87% and testing accuracy of 99.56%. Testing on ECG signals from different time intervals before the onset of SCA demonstrates the model’s robustness and efficiency. The fusion of CNN and LSTM enhances both prediction accuracy and reliability, improving sensitivity and specificity. This approach not only delivers high accuracy but also strengthens the model’s overall predictive power, making it a valuable tool for advanced cardiovascular monitoring systems. Moreover, the model’s low computational complexity enables real-time early detection of Sudden Cardiac Arrest (SCA) with minimal delay, offering a promising solution for proactive cardiac health management.