Sleep apnea is a common sleep disorder that causes repeated interruptions in the breath cycle during sleep, leading to a loss in quality of sleep and an increased risk of cardiovascular diseases. This study presents an approach that uses Deep learning to detect sleep apnea using features extracted from ECG signals. The model development was done using the Apnea-ECG database, and the preprocessing pipeline includes Daubechies-4 wavelet denoising, Z-score-based normalization, and cubic spline interpolation applied to R-R intervals and QRS complexes. These pre-processed signals are used to train and evaluate multiple architectures, including Vanilla LSTM, CNN-LSTM, and 1D convolutional neural networks (CNNs). The results demonstrate the importance of these models in distinguishing between normal and sleep apnea events, showcasing the potential of deep learning methods for real-time sleep apnea detection. The results demonstrate the importance of these models in distinguishing between Sleep apnea and normal events, with our proposed methodology achieving a performance of over 95%. This approach offers a promising solution for improving the diagnosis and management of sleep disorders.