Many elderly individuals prefer to live independently at home for as long as possible. Fall detection systems provide a safety environment, allowing seniors to maintain their independence while having peace of mind. Detecting falls early enables prompt medical interventions, which can prevent further injury or complications, especially in cases where the elderly person is unable to call for help themselves. It can also help elderly individuals to maintain their independence by preventing prolonged hospital stays or long term disabilities resulting from fall related injuries. Enhanced elderly fall detection is a cutting edge solution aimed at revolutionizing safety protocols in healthcare facilities. This paper proposes an enhanced fall detection system leveraging a fusion of Convolutional Neural Network (CNN) and Recurrent Neural Networks (RNN). CNN is used for spatial processing whereas RNN which consists of Gated Recurrent Unit (GRU) layers is used for temporal processing. Experiments conducted on real-world standard data set demonstrate the effectiveness of the CNN- RNN fusion approach and significant improvements in both detection accuracy and false message reduction. Performance of the model is evaluated based on various key metrics highlighting the potential of deep learning fusion frameworks in enhancing elderly fall detection systems.