Chronic Heart Failure (CHF) is a life-threatening cardiovascular condition that requires early and accurate diagnosis to improve patient outcomes. However, conventional diagnostic techniques such as echocardiography are expensive and require specialized medical expertise, limiting their accessibility in resource-constrained settings. This paper proposes a hybrid approach that integrates Machine Learning (ML) and end-to-end Deep Learning (DL) techniques for the automated detection of CHF using phonocardiogram (PCG) signals. Heart sound data is collected from the PhysioNet dataset and preprocessed through noise reduction, normalization, and segmentation of cardiac cycles. Time-domain and frequency-domain features are extracted and used to train a Random Forest classifier, while spectrogram representations are fed into a Convolutional Neural Network (CNN) for automatic feature learning. A hybrid model is then developed by combining the strengths of both approaches to improve classification performance. The system is evaluated using metrics such as accuracy, sensitivity, and specificity, demonstrating enhanced performance in distinguishing normal and CHF conditions. The proposed method provides a non-invasive, cost-effective, and reliable solution that can assist healthcare professionals in early diagnosis and remote monitoring of heart failure.